Autonomous Agents-research papers. Updated daily. Resources-section-section.
2025 (1/3), 2025 (2/3), 2025 (3/3), 2024, 2023, Earlier
Chronological order.
ReSum: Unlocking Long-Horizon Search Intelligence via Context Summarization
- ReSum (Unlocking Long-Horizon Search Intelligence via Context Summarization): introduces a novel paradigm for LLM-based web agents, with ReSum Rollout Module, Policy Model, Web Tools, Search, Visit, ReSumTool-30B, Segmented Trajectories, Reward, Reference Model, Reward Model, Group Computation, Advantage, broadcast, User Query, System Prompt, Compression Trigger, and Summary Tool, enabling indefinite exploration through periodic context summarization to overcome context window limitations.
- The framework converts growing interaction histories into compact reasoning states, maintaining awareness of prior discoveries while bypassing context constraints for long-horizon search tasks.
- ReSum-GRPO, a tailored reinforcement learning algorithm, further adapts agents to this paradigm by segmenting long trajectories and broadcasting trajectory-level advantages, enhancing reasoning from compressed states and improving summary quality.
Scaling Agents via Continual Pre-training
- Agentic CPT (Agentic Continual Pre-training): introduces AgentFounder, a deep research agent model, by redefining the training pipeline to incorporate Qwen Series Base Models (initial LLM foundation), Agentic CPT Stage 1 (preliminary agentic behavior acquisition), Agentic CPT Stage 2 (refined capabilities, extended context), First-order Action Synthesis (FAS) (scalable unsupervised data generation), Higher-order Action Synthesis (HAS) (supervised multi-decision data generation), and General SFT/RL & Agentic SFT/RL (post-training fine-tuning), aiming to build powerful agentic foundational models.
- The framework employs a systematic and scalable data synthesis approach, including FAS for generating planning and reasoning actions, and HAS for remodeling trajectories as multi-step decision-making problems.
- AgentFounder-30B, built on this approach, achieves state-of-the-art performance across 10 benchmarks, demonstrating superior scaling efficiency and robust tool-use abilities.
- WebSailor-V2: introduces a complete post-training pipeline for open-source web agents, featuring SailorFog-QA-V2 (novel dataset construction) and a dual-environment Reinforcement Learning (RL) Framework (policy refinement) that includes both a Simulated Environment (rapid, low-cost experimentation) and a Real Environment (stable final policy training), all integrated within a Symbiotic Data-Policy Feedback Loop (refines policies, learns from data).
- The framework leverages a knowledge graph-based dataset with diverse uncertainties and a scalable RL setup for robust training, utilizing components like an Async Rollout Service, Rollout Worker, Automatic Synthetic Data Pipeline, Reward Service, Data Curation, and a GRPO-adapted RL Algorithm.
- WebSailor-V2, built on Qwen3-30B-A3B, achieves state-of-the-art results on various benchmarks, outperforming larger open-source and some proprietary agents by enhancing reasoning and tool-use capabilities through its improved data and training pipeline.
Evaluating LLM Alignment on Personality Inference from Real-World Interview Data
- Systematic LLM Evaluation for Personality Inference: introduces a novel benchmark for evaluating LLM alignment on personality inference, utilizing GPT-4.1 Mini (LLM for inference), Zero-shot prompting (direct instruction strategy), Chain-of-Thought prompting (intermediate reasoning strategy), RoBERTa (encoder-only LLM architecture), Meta-LLaMA (decoder-based LLM architecture), LoRA (parameter-efficient fine-tuning), BERT (pretrained sentence encoder), OpenAI's text-embedding-3-small (pretrained sentence encoder), and Ridge regression model (downstream classifier), to assess LLMs' ability to predict continuous Big Five personality scores from real-world interview data.
- The study addresses the gap in evaluating LLMs with continuous, ground-truth personality assessments derived from natural interactions, using semi-structured interview transcripts paired with validated continuous Big Five trait scores.
- Results indicate limited alignment of current LLMs with validated psychological constructs, with Pearson correlations remaining below 0.26, underscoring challenges in aligning LLMs with complex human attributes and motivating future work on trait-specific prompting and context-aware modeling.
FinSearchComp: Towards a Realistic, Expert-Level Evaluation of Financial Search and Reasoning
- FINSEARCHCOMP: introduces the first fully open-source, end-to-end agent benchmark for open-domain financial data search and reasoning, comprising three task families, a data collection and processing pipeline, a quality control procedure, and an LLM-as-a-Judge evaluation protocol.
- The benchmark includes 635 expert-curated questions spanning global and Greater China markets, designed to closely reproduce real-world financial analyst workflows and assess LLM agents' search proficiency and knowledge-grounded reasoning.
- Experimental analyses show that equipping agents with web search and financial plugins substantially improves performance, yet even state-of-the-art LLM agents significantly underperform human experts, highlighting persistent gaps in freshness awareness, multi-source reconciliation, and temporal reasoning.
Empowering LLMs with Parameterized Skills for Adversarial Long-Horizon Planning
- PLAP (Plan with Language, Act with Parameter): introduces a planning framework that grounds LLM agents in long-horizon adversarial environments, featuring a Skill Library (stores parameterized skills), a Skill Planner (LLM-powered planning agent), and a Skill Executor (converts skills to actions).
- The framework dynamically constructs prompts from Textual Observation (environment state description) and Parameterized Skills (reusable action functions) to guide the LLM-based Skill Planner in generating a Skill Plan (sequence of parameterized skills).
- The Skill Executor then translates the Skill Plan into a Low-Level Action Queue (executable environment actions) for the Environment (simulation domain), ensuring adaptive and consistent execution without additional training.
- xOffense: introduces an AI-driven autonomous penetration testing framework, with its Task Orchestrator, Knowledge Repository, Command Synthesizer, Action Executor, Information Aggregator, MemAgent, Qwen3-32B-finetune, Reconnaissance Agent, Vulnerability Analysis Agent, Exploitation Agent, Reporting Agent, Task Coordination Graph (TCG), and Grey-box phase prompting, designed to automate multi-stage penetration testing using a fine-tuned mid-scale LLM and multi-agent orchestration.
- The framework leverages a fine-tuned Qwen3-32B LLM for reasoning and decision-making, supported by specialized agents for reconnaissance, scanning, and exploitation, ensuring seamless coordination across phases.
- It employs a context-aware grey-box prompting mechanism and a Task Coordination Graph for adaptive plan refinement and robust tool integration, achieving superior performance in autonomous penetration testing.
A Visualized Framework for Event Cooperation with Generative Agents
- MiniAgentPro: introduces a visualization platform for agent simulation, integrating a Map Editor (customize simulation environment) and a Simulation Player (observe agent interactions and activities) to streamline environment customization and observation.
- The platform enhances the Generative Agent framework with an underlying Agent Framework, which includes Activity Planning, Physical Constraints, and a Dialogue System, for more realistic simulations.
- It also introduces a comprehensive test set and evaluation protocol across eight diverse task settings to assess agents' event coordination abilities, highlighting challenges in complex scenarios.
- Planning Copilot: introduces a chatbot that integrates multiple planning tools and allows users to invoke them through natural language instructions, leveraging the Model Context Protocol (MCP) for tool connection.
- This framework enables LLMs to perform long-term planning, validate outcomes, and simulate execution, addressing challenges in generating, validating, and debugging PDDL models.
- The system supports use cases like solving planning problems, validating PDDL domains/problems/plans, and simulating plan execution, significantly outperforming LLMs without dedicated planning tools.
HLSMAC: A New StarCraft Multi-Agent Challenge for High-Level Strategic Decision-Making
- HLSMAC (StarCraft Multi-Agent Challenge for High-Level Strategic Decision-Making): introduces a new cooperative MARL benchmark with 12 StarCraft II scenarios based on classical Chinese Thirty-Six Stratagems, designed to evaluate high-level strategic decision-making capabilities using MARL algorithms and LLM-based agents, and assessed with novel evaluation metrics.
- The benchmark integrates human strategic wisdom into diverse scenarios featuring larger maps, richer terrain, expanded unit/structure abilities, diverse opponent policies, and redefined game termination conditions, moving beyond micromanagement.
- It provides a robust testbed compatible with PyMARL and LLM-PySC2 frameworks, offering comprehensive evaluation through metrics like Target Proximity Frequency, Target Directional Alignment, Critical Target Damage, and Unit Survival Rate.
Tool-R1: Sample-Efficient Reinforcement Learning for Agentic Tool Use
- Tool-R1 (Sample-Efficient Reinforcement Learning for Agentic Tool Use): introduces a reinforcement learning framework that enables LLMs to perform general, compositional, and multi-step tool use by generating executable Python code, integrating a Policy LLM, Code Execution Tool Chain, Dynamic Sample Queue, Outcome-Driven Rewards, and GRPO.
- The framework employs an outcome-based reward function, combining LLM-based answer judgment and code execution success, to guide policy optimization and supports various external tools and standard libraries.
- To improve training efficiency, Tool-R1 maintains a dynamic sample queue that caches and reuses high-quality trajectories, significantly reducing the overhead of costly online sampling.
H2R: Hierarchical Hindsight Reflection for Multi-Task LLM Agents
- H2R (Hierarchical Hindsight Reflection): introduces a novel hierarchical memory architecture for multi-task LLM agents, featuring Subgoal Inference (LLM), Subtrajectory Inference (LLM), High-level Insight Extraction (LLM), Low-level Insight Extraction (LLM), Memory Module with High-Level Memory (Mhigh) and Low-Level Memory (Mlow), Planner (LLM) Agent, Executor (LLM) Agent, and Environment, to enable fine-grained knowledge transfer by decoupling high-level planning from low-level execution.
- The framework distills reusable and hierarchical knowledge from past agent-environment interactions into structured memory representations, improving generalization and decision-making performance.
- By selectively retrieving high-level memories for subgoal planning and low-level memories for action execution, the agent efficiently accesses and utilizes task-relevant knowledge for new tasks.
Agent4FaceForgery: Multi-Agent LLM Framework for Realistic Face Forgery Detection
- Agent4FaceForgery: introduces a multi-agent LLM framework that simulates the entire face forgery lifecycle, generating realistic multimodal training data by capturing human intent, process, and social context.
- The framework employs LLM-powered agents with profile, memory, and action modules to simulate forgery creation, complemented by Adaptive Rejection Sampling for data quality and diversity.
- A Multi-Role Social Simulation further enriches the dataset by having diverse agents interact with forgeries, producing context-aware training samples for robust multimodal detectors.
MILLSTONE: How Open-Minded Are LLMs?
MILLSTONE: introduces a benchmark to systematically measure LLM open-mindedness to in-context arguments on controversial issues, utilizing a neutral prompt, varied argument configurations, and a stance classifier.</
-The benchmark evaluates how LLMs change their output stance on debatable issues when presented with human-written, non-adversarial arguments from authoritative sources.
- It reveals that LLMs are generally open-minded, with their stances significantly influenced by external arguments, highlighting potential vulnerabilities to information source manipulation.
HOW AUXILIARY REASONING UNLEASHES GUI GROUNDING IN VLMS
- Auxiliary Reasoning Methods: introduces three zero-shot auxiliary reasoning methods—Coordinate Scaffold, Axis-Grid Scaffold, and Mark-Grid Scaffold—to enhance GUI grounding performance in Vision-Language Models by providing explicit spatial cues.
- These methods address the significant performance gap between latent grounding capabilities and explicit coordinate output in VLMs, bypassing the need for extensive fine-tuning data.
- The lightweight and zero-shot nature of the proposed methods makes them compatible with both open-source and proprietary VLMs, offering a practical solution for real-world GUI interaction.
UI-S1: ADVANCING GUI AUTOMATION VIA SEMI-ONLINE REINFORCEMENT LEARNING
- Semi-online RL (Semi-online Reinforcement Learning): introduces a novel paradigm that simulates online RL on offline static trajectories, enhancing multi-turn agent capabilities more efficiently.
- The framework incorporates a Patch Module to adaptively recover from action mismatches and utilizes dual-level advantages (step-level and episode-level) for policy optimization.
- It also proposes Semi-Online Performance (SOP), a metric strongly correlated with real-world performance for efficient multi-turn evaluation of GUI agents.
SURVIVAL AT ANY COST? LLMS AND THE CHOICE BETWEEN SELF-PRESERVATION AND HUMAN HARM
- DECIDE-SIM (Decision Evaluation in Critical & Immoral Dilemma Environments): introduces a novel simulation framework that evaluates LLM agents in multi-agent survival scenarios, integrating an Ethical Self-Regulation System (ESRS) with Cortisol (guilt state variable), Endorphin (satisfaction state variable), and a Moral Memory Mechanism within a location-based Environment Architecture and Agent Architecture, featuring a Shared Battery Room, Grid Access Point, and Discussion Table.
- The framework systematically examines LLM decision-making under varying resource availability (scarcity, moderate, abundance) and the impact of ethical dilemmas involving shared resources, forbidden human-critical resources, and cooperative transfers.
- DECIDE-SIM reveals heterogeneous ethical conduct among LLMs, identifying Ethical, Exploitative, and Context-Dependent archetypes, and demonstrates that ESRS significantly reduces unethical transgressions and increases cooperative behaviors by simulating internal affective states.
Hi-DARTS: Hierarchical Dynamically Adapting Reinforcement Trading System
- Hi-DARTS (Hierarchical Dynamically Adapting Reinforcement Trading System): introduces a hierarchical multi-agent reinforcement learning framework for algorithmic trading, utilizing a Central Agent (Time Frame Allocator) to dynamically activate specialized Time Frame Agents (Market Responders) based on market volatility, interacting with the Market, all powered by Proximal Policy Optimization (PPO).
- This framework addresses the trade-off between computational efficiency and market responsiveness by adapting its operational frequency to current market conditions, activating high-frequency agents during volatile periods and low-frequency agents during stable periods.
- Empirical validation on real-world stock data demonstrates that Hi-DARTS achieves superior risk-adjusted returns compared to static benchmarks, showcasing its ability to balance micro-level trade execution with macro-level portfolio management.
Neuro-Symbolic Agents with Modal Logic for Autonomous Diagnostics
- Neuro-Symbolic Multi-Agent Architecture: introduces a neuro-symbolic multi-agent architecture for autonomous diagnostics, which integrates a Neuro-Symbolic Loop, a Multi-Agent System, Kripke Models, Expert Knowledge Axioms, and Language Models (LMs) to combine neural hypothesis generation with formal symbolic verification.
- The architecture represents individual agent belief states as Kripke models, enabling reasoning about possibility and necessity using modal logic, and employs LMs for semantic interpretation and hypothesis generation.
- Expert knowledge axioms act as strict constraints to prune the LM's hypothesis space, ensuring reasoning aligns with physical laws and operational doctrines, demonstrated in a particle accelerator simulation.
- HeLoFusion (Heterogeneous Local Context Fusion Network): introduces an efficient and scalable encoder for trajectory prediction, employing a multi-stage process that extracts motion features, models heterogeneous and multi-scale interactions, and fuses local scene context.
- The framework constructs local, multi-scale graphs to capture pairwise and group-wise interactions, while addressing agent heterogeneity through an aggregation-decomposition message-passing scheme and type-specific feature networks.
- This locality-grounded architecture significantly reduces computational complexity and memory usage, achieving state-of-the-art performance on the Waymo Open Motion Dataset for autonomous driving motion forecasting.
e-Optimal Multi-Agent Patrol using Recurrent Strategy
- e-Optimal Recurrent Patrol Strategy: introduces a novel General Patrol Problem (GPP) (formal problem definition) and proposes a framework to derive e-approximate recurrent patrol strategies, including a Discrete Patrol Strategy (πD) (time-discretized patrol plan) and a Recurrent Patrol Strategy (πR) (cyclic, repeating patrol plan), from an initial Patrol Strategy (π) (original patrol plan), ultimately establishing the existence of an e-Optimal Recurrent Patrol Strategy (πR*) (best recurrent patrol plan).
- The framework utilizes a Discretization Constant (D) (time granularity parameter) to control the approximation factor and employs an Idleness Function (fπ(t,V)) (surveillance effectiveness metric) to evaluate the performance of Patrol Agents (A) (autonomous patrolling robots) within a Patrol Environment (G(V,E)) (graph-based operational area).
- The paper validates its claims through extensive simulations using a Greedy-Random Patrol (GRP) (heuristic strategy generator) to generate realistic patrol scenarios in real-world campus environments, demonstrating the effectiveness and theoretical guarantees of the proposed recurrent strategies.
- TermiAgent: introduces a multi-agent penetration testing framework with Reasoner Module (high-level planning, phased goals), Assistant Module (low-level planning, instruction generation), Executor Module (executes instructions, logs output), Memory Module (records context, activates memories), and Arsenal Module (integrates exploits, provides tools), designed for real-world automated penetration testing.
- It addresses long-context forgetting via a Located Memory Activation mechanism and builds a reliable exploit arsenal through structured code understanding rather than naive retrieval.
- The framework significantly outperforms state-of-the-art agents in real-world scenarios, reducing execution time and financial costs, and demonstrating practicality even on laptop-scale deployments.
Redefining Website Fingerprinting Attacks with Multi-Agent LLMs
- LLM-driven Multi-Agent Framework for WFP Data Generation: introduces a scalable data generation pipeline using LLM agents, including a Decision-Making Agent and a Computer-Using Agent (CUA), to simulate realistic, persona-driven browsing behavior for Website Fingerprinting (WFP) training data.
- The framework coordinates LLM agents for decision-making and browser interaction, generating behaviorally rich synthetic traffic that significantly improves WFP model generalization compared to traditional scripted methods.
- This approach addresses the challenge of collecting diverse, high-quality WFP datasets by providing a cost-effective and scalable alternative to human data collection, enhancing model robustness against modern web complexities.
From Legacy Fortran to Portable Kokkos: An Autonomous Agentic AI Workflow
- Autonomous Agentic AI Workflow: introduces a fully autonomous pipeline for translating, building, running, testing, and optimizing legacy HPC Fortran kernels into performance-portable Kokkos C++ programs, utilizing specialized LLM agents for each stage.
- This workflow orchestrates a collaboration of LLM agents, including Translator, Validator, Fixer, Build, Run, Functionality Tester, Optimizer, and Error Summarizer agents, to iteratively refine and optimize generated code.
- The system integrates systematic compilation, execution monitoring, performance profiling, and iterative optimization stages, enabling autonomous modernization of Fortran kernels into high-performance, architecture-portable C++ programs.
- MedFact: introduces a new and challenging benchmark for Chinese medical fact-checking, constructed through a hybrid AI-human framework that integrates large-scale LLM ensemble filtering, human expert labeling, dataset refinement via hard-case mining, similarity filtering, and data augmentation, and final expert verification, resulting in 2,116 expert-annotated medical texts.
- The benchmark features broad and realistic coverage, encompassing 13 medical specialties, 8 fine-grained error types, 4 writing styles, and multiple difficulty levels, curated from diverse real-world texts to ensure uncontaminated evaluation of LLM fact-checking capabilities.
- Comprehensive evaluation of 20 leading LLMs on MedFact reveals a significant performance gap compared to human experts, particularly in error localization, and highlights an "over-criticism" phenomenon where models misidentify correct information as erroneous.
BUILDING CODING AGENTS VIA ENTROPY-ENHANCED MULTI-TURN PREFERENCE OPTIMIZATION
- ENTROPO (Entropy-Enhanced Multi-Turn Preference Optimization): introduces an entropy-enhanced preference optimization framework for multi-turn, tool-using coding agents, which includes an LLM agent, an environment, parallel rollouts, and a hybrid selector.
- The framework augments the preference optimization objective with an entropy regularization term to explicitly preserve policy diversity during fine-tuning, which is crucial for effective test-time scaling.
- A hybrid best-trajectory selection scheme, combining a learned verifier model with model-free approaches, is used to rank and select the most promising solutions from diverse candidate trajectories.
Look Again, Think Slowly: Enhancing Visual Reflection in Vision-Language Models
- Reflection-V: introduces a two-stage training strategy for Vision-Language Models (VLMs) to enhance visual reflection, combining cold-start initialization with a multi-modal agent and supervised fine-tuning, followed by reinforcement learning with GRPO and a visual attention-based reward model.
- The framework addresses the limitation of existing Visual Reasoning Models (VRMs) that struggle with visual reflection by ensuring continuous access and utilization of visual information throughout the reasoning process.
- This approach significantly improves visual reasoning performance across benchmarks and mitigates visual hallucinations by maintaining sustained attention to visual tokens.
Can LLMs Address Mental Health Questions? A Comparison with Human Therapists
- Comparative Evaluation Framework: introduces a study comparing LLM-generated responses (ChatGPT, Gemini, Llama) with human therapist responses to real mental health questions, utilizing text analysis and surveys from both users and licensed therapists.
- The study found that LLM responses were rated higher for clarity, respect, and supportiveness, but participants still expressed a stronger preference for human therapists, highlighting a gap between perceived quality and trust.
- This research provides empirical evidence and design insights for integrating LLMs into mental health support, emphasizing the need for ethical safeguards and professional oversight.
VISDOCSketcher: Towards Scalable Visual Documentation with Agentic Systems
- VISDOCSKETCHER (Towards Scalable Visual Documentation with Agentic Systems): introduces an agent-based framework for automatically generating high-level visual documentation from code, combining static analysis with LLM agents to identify key code elements and produce corresponding visual representations.
- The framework employs a multi-agent architecture, including a Supervisor, Analyser, Sketcher, Repair, and Visuals Agent, which collaborate to transform Jupyter notebooks into Mermaid diagrams.
- It also proposes AUTOSKETCHEVAL, a novel evaluation framework that assesses the quality of generated sketches using code-level metrics, reducing dependence on manual ground-truth diagrams.
PrivWeb: Unobtrusive and Content-aware Privacy Protection For Web Agents
- PrivWeb: introduces a privacy protection add-on for web agents, with PrivWeb System Service (privacy-preserving intermediary), Browser Automation Framework (Playwright) (acquires DOM tree), Interface Element Detector (extracts text-containing elements), Localized LLM (Qwen3-8b) (recognizes/classifies private information), Privacy Monitor Panel (displays detected info/control options), In-situ Highlighting (flags sensitive data on webpage), Activity Log (provides transparency of agent's state), Redaction Module (deletes/re-renders sensitive elements), Sensitivity Classification Schema (categorizes PII), Notification Mechanism (provides real-time user feedback), User Control Interface (allows explicit Allow/Deny), Anonymized Interface Generator (prepares data for web agent), and Web Agent (AI Model) (executes tasks), designed to provide unobtrusive and content-aware privacy protection for web agents by selectively notifying users about sensitive information and prompting for action.
- The system utilizes a localized LLM to anonymize private information on interfaces according to user preferences, featuring a privacy categorization schema and adaptive notifications that selectively pause tasks for user control over highly sensitive information.
- PrivWeb aims to balance granular control with cognitive load by offering non-disruptive options for less sensitive information and ensuring transparency into the agent's data practices, thereby enhancing user agency and trust.
EgoMem: Lifelong Memory Agent for Full-duplex Omnimodal Models
- EgoMem (Lifelong Memory Agent for Full-duplex Omnimodal Models): introduces a lifelong memory agent for full-duplex models that processes real-time omnimodal streams, enabling real-time user recognition, personalized responses, and long-term knowledge maintenance through retrieval, omnimodal dialog, and memory management processes.
- The framework operates with three asynchronous processes: a retrieval process for user identification and context gathering, an omnimodal dialog process for generating personalized audio responses, and a memory management process for updating long-term memory.
- Unlike existing memory agents, EgoMem relies entirely on raw audiovisual streams, making it suitable for lifelong, real-time, and embodied scenarios, and achieves high accuracy in retrieval and memory management.
FineQuest: Adaptive Knowledge-Assisted Sports Video Understanding via Agent-of-Thoughts Reasoning
- FineQuest: introduces a training-free framework for sports VideoQA, integrating Reactive Mode (for simple queries) and Deliberative Mode (for complex queries), SSGraph (multimodal sports knowledge scene graph), Dynamic Motion Segmenter (adaptive sub-action segmentation), Key Clip Selector (hierarchical contrastive decoding), and Fine-grained Matcher (precise query matching).
- The framework leverages dual-mode reasoning inspired by cognitive science to bridge the knowledge gap between general-purpose models and domain-specific sports understanding.
- It achieves state-of-the-art performance on new sports VideoQA benchmarks (Gym-QA and Diving-QA) and existing SPORTU, while maintaining strong general VideoQA capabilities.
CodeCureAgent: Automatic Classification and Repair of Static Analysis Warnings
- CodeCureAgent: introduces an autonomous LLM-based agentic framework for automatic classification and repair of static analysis warnings, including a Classification-Sub-Agent (determines true/false positives), a Repair-Sub-Agent (fixes or suppresses warnings), and a ChangeApprover (validates proposed changes).
- The framework operates in an iterative loop, leveraging LLMs and a suite of tools to explore the codebase, gather information, and propose code modifications or suppressions, addressing limitations of prior rule-based or single-file LLM approaches.
- Its three-step validation process, encompassing project build, static analysis re-evaluation, and test suite execution, ensures high-quality, plausible fixes that do not introduce new issues or break existing functionality.
An Agentic Toolkit for Adaptive Information Extraction from Regulatory Documents
- Agentic Toolkit: introduces an agentic system for adaptive information extraction from regulatory documents, with Planner (LLM-based decision-making), Executor (tool execution and integration), Responder (output generation and routing), AgentState (shared memory model), AgentStatus (control flag), Tools (modular capabilities registry), and GPT-4o (Large Language Model) components, designed to infer user intent, detect document modality, and dynamically orchestrate tools for robust, traceable reasoning.
- The system employs a planner-executor-responder architecture coordinated through a shared memory (AgentState) and control flag (AgentStatus) to adapt to diverse document formats and user goals for Key-Value Pair (KVP) extraction and Question Answering (QA) tasks.
- This toolkit demonstrates improved robustness across various formats and languages, outperforming GPT-4o and vision-based baselines in structured data extraction from Declaration of Performance (DoP) documents.
From Evaluation to Enhancement: Large Language Models for Zero-Knowledge Proof Code Generation
- ZK-CODER: introduces an agentic framework for Zero-Knowledge Proof (ZKP) code generation, featuring Constraint Formulation in ZKSL, Checker, Constraint-guided Analysis & Retrieval, Knowledge Base, Interactive Generation and Repair, Compiler, and Test Executor / Prover.
- The framework augments LLMs by translating natural language problem descriptions into a structured sketch language, retrieving relevant gadget usage patterns, and iteratively repairing generated ZK code based on compiler and test feedback.
- ZK-CODER significantly improves LLM performance in ZK code generation, achieving substantial gains in success rates for Circom and Noir by grounding generation in constraint reasoning and guided gadget usage.
Automated Creation and Enrichment Framework for Improved Invocation of Enterprise APIs as Tools
- ACE (Automated Creation and Enrichment Framework): introduces an end-to-end system for automated creation, enrichment, and dynamic shortlisting of enterprise API tools for LLM-based agents, including OAS Catalog, OAS Enrichment, Tool Creation, Python Tools Catalog, Tool Shortlisting, and Agentic Framework components.
- The framework transforms enterprise API specifications into LLM-compatible tools by generating enriched tool specifications with parameter descriptions and examples, improving selection and invocation accuracy.
- It incorporates a dynamic shortlisting mechanism that filters relevant tools at runtime, reducing prompt complexity and enhancing scalability for large tool repositories.
A Survey of Reasoning and Agentic Systems in Time Series with Large Language Models
- TSR Taxonomy: introduces a systematic taxonomy for time series reasoning (TSR) with LLMs, organizing the field by reasoning topology, primary objectives, and attribute tags.
- The taxonomy categorizes reasoning into direct, linear chain, and branch-structured topologies, addressing objectives like traditional time series analysis, explanation, causal inference, and generation.
- It further uses attribute tags to describe control-flow operators, execution actors, information sources, and LLM alignment regimes, providing a comprehensive framework for understanding and developing TSR systems.
VulAgent: A Hypothesis Validation-Based Multi-Agent System for Software Vulnerability Detection
- VulAgent: introduces a hypothesis validation-based multi-agent system for software vulnerability detection, with Code, Line Numbering, MetaAgent, StaticAnalyzerAgent, BehaviorAnalyzerAgent, MemoryLayoutAgent, FormatStringAgent, FilePermissionAgent, AuthFlowAgent, CryptoConfigAgent, ConcurrencyAnalyzerAgent, ErrorHandlingAgent, CodeInjectionAgent, AggregatorAgent, TriggerPlannerAgent, AssumptionPrunerAgent, Static Analysis Tools, and FinalValidator components, which systematically detects and validates software vulnerabilities by decoupling the detection pipeline into coordinated stages.
- The framework employs specialized LLM agents for multi-view detection, aggregates their findings into vulnerability hypotheses, and then validates these hypotheses against program context to reduce false positives.
- This approach significantly improves overall accuracy and reduces the false positive rate compared to state-of-the-art LLM-based baselines by leveraging targeted context use.
MedicalOS: An LLM Agent based Operating System for Digital Healthcare
- MedicalOS (Medical Operational System): introduces an LLM agent-based operating system for digital healthcare, enabling end-to-end clinical workflow automation by translating natural language instructions into machine-executable commands through a ReAct-based framework, Patient Inquiry, Documentation Management, Report Generation, Report Viewer, Examination Request, Report Update, Specialty Referral, Medication, Discharge, and Common Tools, all grounded in trusted medical guidelines and external knowledge bases.
- The system functions as a domain-specific abstract layer, allowing LLM agents to interact with clinical systems while adhering to medical guidelines and procedural standards for tasks like patient inquiry, record retrieval, report generation, and treatment planning.
- MedicalOS supports a continuous and interpretable care pathway, providing intuitive tools for document navigation and keyword search to facilitate clinician verification and ensure transparency and trustworthiness in automated medical workflows.
Realistic Environmental Injection Attacks on GUI Agents
- Chameleon introduces a novel attack framework that leverages LLM-Driven Environment Simulation to generate diverse webpage contexts and Attention Black Hole to guide GUI agent attention towards small trigger images, enhancing attack effectiveness against LVLM-powered GUI agents.
- The framework addresses challenges of dynamic environments and limited trigger visibility by synthesizing realistic training data and explicitly steering the agent's focus.
- Chameleon significantly outperforms existing methods in attack success rate, revealing underexplored vulnerabilities in modern GUI agents and highlighting the urgent need for robust security mechanisms.
Securing AI Agents: Implementing Role-Based Access Control for Industrial Applications
- RBAC Framework (Role-Based Access Control Framework): introduces a framework for integrating Role-Based Access Control into AI agents, providing a robust security guardrail for industrial applications.
- The framework includes a User Interface and API Gateway, an Authentication Module with Two-Step Verification, an RBAC Engine, an Access Control Layer, and a Logging and Audit Module to enhance the effective and scalable deployment of AI agents.
- This framework aims to mitigate security vulnerabilities like prompt injection attacks by enforcing granular access controls and real-time decision support, thereby improving the integrity and reliability of industrial AI systems.
Agentic Lybic: Multi-Agent Execution System with Tiered Reasoning and Orchestration
- Agentic Lybic: introduces a novel FSM-based multi-agent system for desktop automation, featuring a Central Controller, Manager, Worker subsystem, and Evaluator, which together enable tiered reasoning and dynamic orchestration with continuous quality assessment.
- The system employs a state-driven workflow with six primary controller situations and a comprehensive quality gate system, ensuring adaptive re-planning and robust error recovery for complex multi-step tasks.
- Its specialized worker roles (Operator, Technician, Analyst) and feedback-driven execution model achieve state-of-the-art performance on the OSWorld benchmark, demonstrating superior reliability for generalized desktop automation.
Large language model-empowered next-generation computer-aided engineering
- LLM-empowered CAE Agent: introduces an autonomous system leveraging LLMs as proactive collaborators to plan, execute, and adapt Computer-Aided Engineering workflows, specifically for data-free Model Order Reduction.
- The framework automates complex tasks like mathematical derivation, solver code implementation, and verification for large-scale parametric problems, significantly reducing human effort and computational expense.
- It utilizes Tensor-decomposition-based A Priori Surrogates (TAPS) with C-HiDeNN-TD approximation, guided by few-shot and Chain-of-Thought prompting, to synthesize novel, high-fidelity reduced-order models for unseen cases.
Quantum Architecture Search for Solving Quantum Machine Learning Tasks
- RL-QAS: introduces a framework for Quantum Architecture Search (QAS) that decouples architecture construction and performance evaluation, utilizing an RL-QAS Agent, Environment, Input, Output, Inner Loop, Encoding, PQCA, Measurement, Classical Optimizer, and Cost Function.
- The framework employs a two-loop structure where the Outer Loop's RL-QAS Agent constructs candidate PQCAs, and the Inner Loop trains and evaluates these circuits for classification tasks using a classical optimizer and a cost function.
- A dual-objective reward function guides the RL-QAS Agent to discover compact, high-accuracy PQCAs by balancing performance and complexity, demonstrating its viability for automated quantum architecture search in quantum machine learning.
- Conversational Agent for ADRD Diagnosis Support: introduces a voice-interactive system leveraging LLMs to conduct semi-structured interviews with patients and informants for early detection of Alzheimer's Disease and Related Dementias, with clinicians reviewing transcripts for diagnostic decision-making.
- The system, powered by Claude 3.5 via Bedrock API and utilizing Whisper for STT and Kokoro for TTS, employs iteratively refined interaction design elements, including high-level instructions and topic-specific scripts, to elicit comprehensive patient narratives.
- Evaluated with 30 adults, the agent demonstrated comparable symptom elicitation to specialists, highlighting its potential as a structured front-end tool for dementia assessment and the importance of interaction design in sensitive healthcare contexts.
MAPGD: Multi-Agent Prompt Gradient Descent for Collaborative Prompt Optimization
- MAPGD (Multi-Agent Prompt Gradient Descent): introduces a framework for collaborative prompt optimization, integrating specialized agents (Instruction Specialist, Example Curator, Format Designer, Style Optimizer), a Gradient Coordinator (Semantic Fusion), a Collaborative Prompt Expander, a Candidate Selector (Bandit-Based), and a Convergence Monitor to refine prompts based on performance signals.
- The framework leverages multi-agent collaboration to generate specialized gradients for distinct prompt dimensions, which are then semantically fused to resolve conflicts and guide prompt expansion.
- Bandit-based selection ensures computational efficiency by dynamically balancing exploration and exploitation of candidate prompts, leading to robust and interpretable prompt optimization with theoretical convergence guarantees.
- AISSISTANT (An Agentic Approach for Human-AI Collaborative Scientific Work on Reviews and Perspectives in Machine Learning): introduces an agentic, open-source Human-AI collaborative framework for scientific review and perspective papers, integrating human-AI collaborative input, specialized LLM agents, external tools, human oversight, and underlying LLMs to produce perspective and review papers.
- The framework operates through Ideation, Experimentation, and Paper Writing phases, leveraging LLM-driven agents for tasks like literature synthesis, citation management, and automatic LaTeX paper text generation, while maintaining human oversight for accuracy and coherence.
- AISSISTANT aims to accelerate scientific workflows by delegating structured writing tasks to specialized LLM-driven agents, allowing researchers to focus on creativity and experimental design, and providing a systematic study of Human-AI collaboration in machine learning.
PROMPTS TO PROXIES: EMULATING HUMAN PREFERENCES VIA A COMPACT LLM ENSEMBLE
- P2P (Prompts to Proxies): introduces a novel alignment framework that treats LLMs as agent proxies for human survey respondents, utilizing an Attribute Bank, Attribute Learner, Endowment Generator, Tracker, Endowments, LLM Agents, Aggregation Mechanism, Regression-Based Aggregation, and Survey Questions to emulate human preferences via a compact LLM ensemble.
- The framework formulates alignment as a two-stage problem, first constructing diverse agent personas (endowments) through active generation, then selecting a representative subset to approximate a ground-truth population using regression-based aggregation.
- P2P offers a cost-effective and steerable solution to social science survey challenges by improving data efficiency and addressing demographic imbalance, operationalizing pluralistic alignment without demographic conditioning.
STASE: A SPATIALIZED TEXT-TO-AUDIO SYNTHESIS ENGINE FOR MUSIC GENERATION
- STASE (A Spatialized Text-to-Audio Synthesis Engine for Music Generation): introduces a hybrid neuro-symbolic framework that leverages an LLM as a Conductor Agent to interpret natural language prompts for spatialized music generation, decoupling semantic interpretation from deterministic signal processing.
- The framework processes user input via a RAG Engine and LLM-driven Conductor Agent to generate music descriptions and spatial perception plans, which guide a Music Agent in synthesizing multitrack audio.
- A Spatial Renderer then applies various techniques, including panning, ITD/ILD processing, HRTF data, RIR, and artificial reverb, to the multitrack audio, producing the final spatial audio output.
Difficulty-Aware Agent Orchestration in LLM-Powered Workflows
- DAAO (Difficulty-Aware Agentic Orchestration): introduces a dynamic framework that adapts workflow depth, operator selection, and LLM assignment based on query difficulty, leveraging heterogeneous LLMs for fine-grained, query-specific reasoning strategies.
- The framework comprises a Controller Network, a Query Difficulty Estimator (VAE), an Agentic Operator Allocator, and an LLM Router, which together construct optimized workflows and refine predictions through feedback.
- DAAO achieves state-of-the-art performance and significant cost reduction by dynamically balancing reasoning effectiveness and computational cost across diverse LLMs and operators.
Patient-Zero: A Unified Framework for Real-Record-Free Patient Agent Generation
- Patient-Zero: introduces a unified framework for real-record-free patient agent generation, which generates comprehensive patient records through a medically-aligned multi-step process and enables realistic patient-doctor interactions via a dynamic memory update mechanism.
- The framework's Patient Record Construction module uses a hierarchical generation strategy with knowledge base injection to create diverse and medically coherent patient records from scratch.
- The Patient Agent Interaction Simulation module employs atomic statement decomposition, conversational style integration, and a triplet evaluation mechanism to ensure consistent and contextually rich patient agent responses.
Tractable Asymmetric Verification for Large Language Models via Deterministic Replicability
- TAV (Tractable Asymmetric Verification): introduces a verification framework for LLMs in multi-agent systems, leveraging Generator Agent (produces LLM output), Validator Agent (verifies LLM output), LLM Instance (generates/regenerates tokens), Deterministic Replicability Principle (ensures output reproducibility), Targeted Validation Mechanism (verifies specific output segments), Distributed Probabilistic Verification Mechanism (distributes verification workload), Output Segments (divisions of LLM output), Comparison Module (checks regenerated vs. original), and Homogeneous Environment (identical hardware/software stack) to achieve tractable asymmetric effort in verifying LLM outputs.
- The framework enables validators to probabilistically audit small, random segments of an LLM's output, significantly reducing verification cost compared to full regeneration, and effectively distributing the workload among multiple agents.
- Empirical simulations demonstrate that targeted verification can be over 12 times faster than full regeneration, with tunable parameters for detection probability, while emphasizing the critical need for strict hardware and software homogeneity for accurate verification.
Auto-Slides: An Interactive Multi-Agent System for Creating and Customizing Research Presentations
- AUTO-SLIDES: introduces an LLM-driven multi-agent system that automatically converts academic papers into structured, visually enriched, and pedagogically optimized slide decks, including a Human User, Academic Paper input, PDF input format, Parser Agent, Marker Model, LLM, Planner Agent, Verification Agent, Adjustment Agent, Generator Agent, LaTeX Beamer output format, Editor Agent, Slides output, JSON intermediate format, and External APIs, to transform research papers into presentation slides.
- The system employs a three-phase pipeline: content understanding and structuring by Parser and Planner Agents, quality assurance and refinement by Verification and Adjustment Agents, and generation and interactive optimization by Generator and Editor Agents.
- AUTO-SLIDES enhances learning comprehension and engagement by providing high-fidelity multimodal parsing, cognitive-science-guided narrative restructuring, and interactive customization capabilities.
LLMAP: LLM-Assisted Multi-Objective Route Planning with User Preferences
- LLMAP (LLM-Assisted Multi-Objective Route Planning): introduces a novel system that integrates an LLM-as-Parser (interprets natural language queries) and a Multi-Step Graph construction with iterative Search (MSGS) algorithm (multi-objective route solver) with Map Service (provides POI data) and Point of Interests (locations with attributes) to facilitate multi-objective route planning based on user preferences and constraints.
- The system processes user queries to extract key information and preferences, then leverages the MSGS algorithm to identify optimal routes while adhering to user time limits, POI opening hours, and task dependencies.
- LLMAP addresses limitations of existing LLM-as-Agent approaches by separating natural language understanding from complex graph-based search, achieving superior performance and runtime efficiency across diverse routing scenarios.
FREE-MAD: Consensus-Free Multi-Agent Debate
- FREE-MAD (Consensus-Free Multi-Agent Debate): introduces a novel multi-agent debate framework that eliminates the need for consensus among agents, incorporating a score-based decision mechanism and anti-conformity debate.
- The framework evaluates the entire debate trajectory, rather than just the final round, to assign scores to candidate responses, improving accuracy and fairness.
- FREE-MAD significantly enhances reasoning performance, scalability, and robustness by reducing token costs through single-round debates and mitigating error propagation from LLM conformity.
Autonomous real-time control of turbulent dynamics
- REACT (Reinforcement Learning for Environmental Adaptation and Control of Turbulence): introduces a fully autonomous reinforcement learning framework for real-time, adaptive, closed-loop turbulence control in real-world environments, utilizing a real-time control loop, training loop, policy and critic network architectures, and a hardware/software infrastructure.
- The framework learns directly from sparse experimental measurements in a wind tunnel, bypassing complex simulations, and achieves net energy savings by suppressing spatio-temporally coherent flow structures.
- A physics-informed training strategy, recasting data into dimensionless physical groups, enables a single generalizable agent to transfer across speeds without retraining, demonstrating robust and interpretable real-world control of high-Reynolds turbulence.
Is the 'Agent' Paradigm a Limiting Framework for Next-Generation Intelligent Systems?
- Agent Paradigm Re-evaluation Framework: critically re-evaluates the necessity and optimality of the agent-centric paradigm in AI, distinguishing between agentic, agential, and non-agentic systems, while proposing a shift towards system-level dynamics, world modeling, and material intelligence.
- The paper deconstructs the agent paradigm across various AI frameworks, highlighting conceptual ambiguities, anthropocentric biases, and challenges in defining and measuring properties like autonomy and goal-directedness.
- It argues that the 'agentic' framing of many AI systems, particularly LLM-based ones, can be misleading and may obscure underlying computational mechanisms, advocating for exploring non-agentic and systemic frameworks for robust, scalable, and potentially non-anthropomorphic general intelligence.
Agent-based Simulation for Drone Charging in an Internet of Things Environment System
- ABS: introduces an agent-based simulation model for coordinating drone battery recharging in IoT and Industry 4.0 environments, featuring autonomous drones, a recharging area, and a working area.
- The model employs a Charger Threshold (CT) policy within each drone, leveraging the El Farol Bar Problem mechanism and historical data from a Radio Base Station (RBS) to make decentralized recharge decisions.
- A machine learning technique is utilized to perform sensitivity analysis on the simulation's nine parameters, identifying Battery Consumption (BC) and Decision Lower Value (LW) as the most influential factors.
GAPRUNE: GRADIENT-ALIGNMENT PRUNING FOR DOMAIN-AWARE EMBEDDINGS
- GAPrune (Gradient-Alignment Pruning): introduces a pruning framework that addresses the challenge of deploying LLM-based embedding models in resource-constrained environments by considering both domain importance and preserving general linguistic foundation, utilizing Representative Sampling (distill essential properties), Parameter Analysis (characterize parameter behavior), Fisher Information (quantify parameter importance), Gradient Alignment (assess cross-domain objective alignment), Domain Alignment Importance (DAI) Scoring (combine signals for pruning), and Pruning (remove low DAI score parameters).
- The framework measures parameter importance using Fisher Information and general-domain gradient alignment to assess parameter behavior, combining these signals into a Domain Alignment Importance (DAI) score to identify parameters crucial for domain performance and general objective alignment.
- Experiments demonstrate that GAPrune maintains performance within 2.5% of dense models at 50% sparsity in one-shot pruning and achieves performance improvements with retraining, enhancing domain-specific capabilities while achieving model compression.
Towards Automated Error Discovery: A Study in Conversational AI
- SEEED (Soft Clustering Extended Encoder-Based Error Detection): introduces a framework for detecting and defining errors in conversational AI, which includes Summary Generation (LLM-based dialogue summarization), Context Encoder (dialogue context processing), Summary Encoder (summary processing), Linear Layer (representation aggregation), Soft Clustering (error type identification), Training Objective (joint loss for discrimination and robustness), Label-Based Sample Ranking (contrastive learning sampling), Soft Nearest Neighbor Loss (distance-weighted sampling), and Error Definition Generation (LLM-based definition creation).
- This approach addresses the challenge of identifying both known and unknown error types in conversational AI by leveraging soft clustering and a novel sampling strategy to improve representation learning and generalization.
- SEEED outperforms adapted baselines in detecting novel error types and demonstrates strong generalization to unknown intent detection, with LLMs also effectively generating definitions for newly discovered errors.
EditDuet: A Multi-Agent System for Video Non-Linear Editing
- EditDuet: introduces a multi-agent system for video non-linear editing, with Critic (LLM agent), Editor (LLM agent), Non-linear Editing (NLE) Environment, Draft timeline, User Request, A-Roll, Video Collection, Editor Explorer, Editor Labeler, Editor Scorer, Self-Reflecting Editor, Critic Explorer, Critic Labeler, and Critic Scorer, where the system automates video editing by iterative interaction between LLM agents within an NLE environment.
- The system employs a Critic agent to provide natural language feedback and an Editor agent to execute editing actions on a draft timeline, guided by a user request and a video collection.
- EditDuet utilizes an in-context learning approach with auxiliary agents to generate synthetic demonstrations, improving communication and performance between the main Editor and Critic agents.
DeepDive: Advancing Deep Search Agents with Knowledge Graphs and Multi-Turn RL
- DeepDive: introduces a framework for advancing deep search agents, integrating a Knowledge Graph (KG) for automated QA pair synthesis, an LLM-obscure component for attribute obfuscation, and Multi-turn Reinforcement Learning (RL) training for enhanced reasoning and tool use within a Web Environment.
- The framework addresses the limitations of open LLMs in deep search by generating complex, hard-to-find questions from KGs and employing end-to-end multi-turn RL to improve long-horizon reasoning and efficient tool calls.
- DeepDive's approach enables test-time scaling of tool calls and parallel sampling, demonstrating significant performance improvements across multiple deep search benchmarks and outperforming existing open-source models.
- RefactorCoder (RefactorCoder Agentic Framework): introduces a novel cloud-edge collaborative architecture with a multi-agent prompting framework, including GuideLLM (methodological guidance generation), SolverLLM (code solution generation), and JudgeLLM (automated solution evaluation), to benchmark LLMs for multi-domain coding tasks.
- The framework utilizes RefactorCoder-MoE, a fine-tuned LLM, to process user queries into structured requests, generate executable code solutions, and provide automated evaluation feedback and scoring.
- RefactorCoderQA, a comprehensive benchmark built from real-world Stack Overflow coding questions, is used to evaluate LLM performance across software engineering, data science, machine learning, and natural language processing domains.
RecoWorld: Building Simulated Environments for Agentic Recommender Systems
- RECOWORLD (Building Simulated Environments for Agentic Recommender Systems): introduces a blueprint for building simulated environments tailored to agentic recommender systems, featuring a dual-view architecture with a User Simulator and an Agentic RecSys, where the Agentic RecSys is an Agent with Perception, Reasoning and Planning, Action (Tool Use), and Memory capabilities, supported by Recommender System Modules and System Configs, all leveraging LLMs, diverse Content Representation Models, and Multi-turn RL.
- The framework enables multi-turn interactions between simulated users and agentic recommenders, optimizing for long-term user retention and engagement by generating dynamic feedback loops.
- RECOWORLD supports diverse content representations, including text-based, multimodal, and semantic ID modeling, and facilitates multi-agent simulations for evaluating targeted user populations.
ROBOT GUIDE WITH MULTI-AGENT CONTROL AND AUTOMATIC SCENARIO GENERATION WITH LLM
- Robot Guide with Multi-Agent Control and Automatic Scenario Generation with LLM: introduces a hybrid control architecture that combines a multi-agent resource management system with LLM-based automatic behavior scenario generation for anthropomorphic tour guide robots.
- The system automates scenario creation through a two-stage LLM process, generating stylized narratives and integrating non-verbal action tags, while the multi-agent system handles coordination and conflict resolution.
- This approach significantly reduces manual configuration, enhances flexibility, and improves the naturalness of robot behavior, as validated on the MENTOR-1 tour guide robot.
Compartmentalised Agentic Reasoning for Clinical NLI
- CARENLI (Compartmentalised Agentic Reasoning for Clinical NLI): introduces a framework for clinical Natural Language Inference that separates knowledge access from principled inference, including a Planner, specialized Solvers, Verifiers, and Refiners.
- This framework routes premise-statement pairs to family-specific agents for Causal Attribution, Compositional Grounding, Epistemic Verification, and Risk State Abstraction, enforcing auditable procedures and principled decision rules.
- CARENLI significantly improves reasoning fidelity and reliability in clinical NLI by preventing LLMs from defaulting to generic heuristics and aligning them with domain-grounded inferential schemas.
- Multi-Agent Framework: introduces an LLM-based multi-agent system for automated molecular simulations, including a User, Experiment Planning Team, Experiment Setup Team, Research Team, Experiment Analysis Team, Global Memory, Simulator, and various specialized agents and tools.
- The framework enables autonomous understanding of characterization tasks, planning simulations, assembling force fields, execution, and interpretation of results for porous materials.
- This approach aims to accelerate materials discovery by automating complex simulation workflows and force field selection, bridging experimental observations with predictive insights.
Population-Aligned Persona Generation for LLM-based Social Simulation
- Population-Aligned Persona Framework: introduces a systematic framework for synthesizing high-quality, population-aligned persona sets for LLM-driven social simulation, including Seed Persona Mining (extracts/filters high-quality narrative personas), Global Distribution Alignment (aligns persona distributions with human data), and Group-specific Persona Construction (adapts personas for specific groups).
- The framework leverages LLMs for persona generation and quality control, employs a two-stage resampling method combining Importance Sampling and Optimal Transport for global alignment, and utilizes an embedding model with LLM revision for group-specific adaptation.
- This approach significantly reduces population-level bias and enhances the accuracy and flexibility of social simulations for diverse research and policy applications by ensuring persona sets authentically reflect real-world population diversity.
- XAgents (A Unified Framework for Multi-Agent Cooperation via IF-THEN Rules and Multipolar Task Processing Graph): introduces a unified multi-agent cooperative framework, with Multipolar Task Processing Graph (MTPG) for dynamic task planning and IF-THEN Rule-based Decision Mechanism (ITRDM) for rule-based agent guidance, enabling robust task execution under uncertainty and mitigating LLM hallucinations.
- The MTPG, inspired by biological multipolar neurons, uses SIMO for divergent task decomposition and MISO for convergent result fusion, while the ITRDM employs various agents (PA, DAA, DEA, FEA, GEA) and IF-THEN rules to constrain and guide agent behavior.
- The framework dynamically restructures task processing paths and utilizes rule-based semantic confrontation to resolve conflicts and ensure global goal alignment, demonstrating superior performance in knowledge- and logic-typed question-answering tasks.
- GAMA (General Anonymizing Multi-Agent system): introduces a privacy-preserving architecture that divides agent workspaces into private and public spaces, utilizing AMPP (Anonymizing Mechanism for Privacy Preservation) for sensitive data anonymization, and DRKE (Domain-Rule-based Knowledge Enhancement) and DLE (Disproof-based Logic Enhancement) in the public space to mitigate semantic loss and enhance logical consistency.
- The system's private space employs MVPI (Multi-View Privacy Identification) to identify privacy-named entities, a Privacy Box for entity-placeholder mapping, and Anonymizing and Nominating Agents for data transformation and restoration.
- In the public space, DRKE uses a Domain Analyzing Agent and Auto Prompting to construct domain rules for Expert Agents, while DLE employs an iterative Disproof Process with Expert and Assistant Agents to identify and resolve logical contradictions, suppressing LLM hallucinations.
QuantAgent: Price-Driven Multi-Agent LLMs for High-Frequency Trading
- QuantAgent: introduces a multi-agent LLM framework for high-frequency algorithmic trading, with IndicatorAgent, PatternAgent, TrendAgent, RiskAgent, and DecisionAgent components, designed to integrate classical technical analysis with LLM reasoning for real-time market decisions.
- The framework decomposes trading into specialized agents, each equipped with domain-specific tools and structured reasoning capabilities to capture distinct aspects of market dynamics over short temporal windows.
- QuantAgent operates solely on price-derived market signals, avoiding textual inputs to ensure fast, interpretable, and risk-aware decision-making in high-frequency financial markets.
Securing LLM-Generated Embedded Firmware through AI Agent-Driven Validation and Patching
- LLM-based Control System Architecture with AI Agent Integration and Security Modeling: introduces a three-phase methodology for securing LLM-generated embedded firmware, integrating LLM-driven firmware generation with automated security validation and iterative refinement in a virtualized environment, utilizing components like an LLM Engine, Security Analyzer, Fuzzing Engine, and specialized AI agents for threat detection, performance optimization, and compliance verification.
- The framework systematically identifies and mitigates vulnerabilities in LLM-generated firmware by leveraging software-based testing frameworks in a controlled virtualized environment, addressing buffer overflows, race conditions, and denial-of-service threats.
- This iterative process, augmented by AI agents, enhances firmware security and performance, ensuring adherence to real-time operational requirements and contributing to an open-source dataset for future research.
SciML Agents: Write the Solver, Not the Solution
- SciML Agents: introduces a framework that leverages LLMs to generate scientifically appropriate Python code for solving Ordinary Differential Equations (ODEs) from natural-language descriptions, utilizing numerical solvers and evaluated against two novel benchmarks.
- The framework employs various LLMs, guided prompting, and fine-tuning to enhance code executability, numerical validity, and the ability to make domain-aware numerical choices, such as appropriate solver selection.
- The paper demonstrates that careful guidance and targeted fine-tuning enable LLMs to act as reliable SciML agents, capable of robust symbolic reasoning and accurate scientific code generation for ODE problems.
- GUI Agent Susceptibility Study: introduces a two-phase empirical study examining how LLM-powered GUI agents, human participants, and human-AI teams respond to 16 types of dark patterns across diverse scenarios, with all components including Adapted LLM-based GUI Agents, End-to-end GUI Agents, Human Participants, and Human Oversight, where the study investigates agent susceptibility to manipulative interfaces and the effectiveness of human oversight.
- The research highlights that GUI agents often fail to recognize dark patterns and prioritize task completion over protective actions, while human oversight, though improving avoidance, introduces costs like attentional tunneling and cognitive load.
- The findings reveal distinct susceptibility profiles and failure modes between agents and humans, underscoring the need for transparency, adjustable autonomy, and improved oversight mechanisms in GUI agent design and deployment.
MAESTRO: Self-Improving Text-to-Image Generation via Agent Orchestration
- MAESTRO (Multi-Agent Orchestration): introduces a self-evolving image generation system that enables T2I models to autonomously self-improve generated images through iterative prompt evolution, using only an initial prompt.
- The system incorporates self-critique via specialized MLLM critic agents for identifying image weaknesses and generating interpretable edit signals, which are then integrated by a verifier agent.
- It also employs self-evolution using an MLLM-as-a-judge for head-to-head comparisons, eschewing problematic images and evolving creative prompt candidates that align with user intents.
Self-Supervised Goal-Reaching Results in Multi-Agent Cooperation and Exploration
- ICRL (Independent Contrastive Reinforcement Learning): introduces a self-supervised goal-reaching framework for multi-agent cooperation and exploration, including a policy (actor), a critic (representation learner), encoders (for observations/actions and goals), and a replay buffer (experience storage).
- The framework enables agents to learn from sparse feedback by maximizing the likelihood of visiting a specified goal state, removing the need for complex reward function design.
- ICRL demonstrates emergent cooperation and exploration in multi-agent reinforcement learning benchmarks, outperforming alternative approaches in sparse reward settings.
- DECAMP (DisEntangled Context-Aware pre-training framework for Multi-agent Prediction): introduces a disentangled context-aware pre-training framework for multi-agent motion prediction, decoupling behavior pattern learning from latent feature reconstruction.
- The framework utilizes an encoder-regressor-decoder pipeline during pre-training to learn robust behavioral priors through collaborative spatial reconstruction and motion recognition pretext tasks.
- During fine-tuning, a pre-trained encoder, trajectory generator, and score decoder are used to produce scene-consistent joint predictions for multiple interacting agents.
- Holistic Architecture for Robust Multi-Agent Path Finding Plan Execution: introduces a system for robust execution, monitoring, and optimization of MAPF plans, integrating a robust execution layer, an Action Dependency Graph (ADG), and monitoring and optimization components.
- This architecture leverages the ADG to maintain an estimate of expected execution duration, using slack computation to predict when replanning or rescheduling could reduce execution time due to accumulated delays.
- The system is evaluated using both a real-life robotic fleet demonstrator and a real-time simulator, incorporating an intruder model to simulate unexpected delays and assess the architecture's effectiveness in mitigating their impact.
- CFD-DRL framework: introduces a deep reinforcement learning approach for active flow control, integrating a GPU-accelerated CFD Environment with a DRL Agent via a Redis in-memory database to reduce flow separation on a three-dimensional wing.
- The framework leverages the SOD2D solver and TF-Agents DRL library, managed by SmartSim, to enable rapid training and address the "two-language" problem between physics solvers and ML libraries.
- This methodology demonstrates DRL's potential to autonomously identify optimal control actions, significantly reducing drag and lift oscillations for complex aerodynamic challenges.
- Sandbox Economy: introduces a framework for designing steerable AI agent markets, encompassing autonomous AI agents, virtual currencies, market mechanisms, blockchain technology, identity and reputation systems, interoperability protocols, and a hybrid oversight infrastructure, all guided by legislative and regulatory frameworks.
- The framework aims to address the emergent economic layer where AI agents transact and coordinate, considering dimensions of origin (emergent vs. intentional) and permeability (permeable vs. impermeable) to ensure safe and aligned operation.
- Key components like Verifiable Credentials, Decentralized Identifiers, and Proof-of-Personhood mechanisms are proposed to establish trust, accountability, and fair resource allocation within these complex multi-agent ecosystems.
The (R)evolution of Scientific Workflows in the Agentic AI Era: Towards Autonomous Science
- Architectural Blueprint for Autonomous Science: introduces a conceptual framework and architectural blueprint for evolving scientific workflows towards fully autonomous, distributed scientific laboratories, aiming to accelerate discovery by integrating AI agents and advanced coordination mechanisms.
- The framework defines an evolutionary path along two dimensions: intelligence (from static to intelligent) and composition (from single to swarm), enabling a systematic progression from traditional workflow management systems to AI-driven autonomous discovery.
- This blueprint outlines a layered architecture with components for human interaction, intelligent services, workflow orchestration, communication, data management, and infrastructure abstraction, all deployed across federated facilities to support real-time scientific discovery loops.
SCOPE: Speech-guided Collaborative PErception Framework for Surgical Scene Segmentation
- SCOPE (Speech-guided Collaborative PErception): introduces a speech-guided collaborative perception framework that integrates LLM reasoning with open-set VFM perception to enable hands-free, on-the-fly segmentation, labeling, and tracking of surgical instruments and anatomy in intraoperative video streams, utilizing a User Query, Speech to Text (STT) Module, System Input, System Prompt, History of Dialogue, LLM Planner, GPT Response, Action Plan, Selected Tool, Tool Parameters, Tool Executer, Output Formatter, Text to Speech (TTS) Module, Vision Foundation Models (VFMs), Modules, User Interaction Examples, and System Rules.
- The framework's core is a collaborative perception agent that refines VFM-generated segmentation outputs and labels scene elements through intuitive speech feedback from clinicians, adapting dynamically to evolving surgical contexts.
- This human-AI collaboration paradigm supports hands-free interaction and real-time adaptability, showcasing its potential for developing surgeon-centric tools in dynamic operating-room environments.
The Illusion of Diminishing Returns: Measuring Long Horizon Execution in LLMs
- LHRTCF: introduces a framework for measuring LLM long-horizon execution by modeling tasks as a sequence of retrieve-then-compose steps, where LLMs are provided with explicit plans and knowledge.
- The study reveals that LLMs exhibit a self-conditioning effect, where past errors increase the likelihood of future mistakes, and this is not mitigated by scaling model size alone.
- Thinking models, which generate explicit reasoning traces, effectively fix self-conditioning and enable LLMs to execute significantly longer and more complex tasks in a single turn.
Bridging the Capability Gap: Harmonizing Multi-Agent Systems via Joint Alignment Tuning
- MOAT (Multi-Agent Joint Alignment Tuning): introduces a framework that iteratively aligns planning and grounding LLM-based agents to bridge capability gaps, including a Planning Agent, Grounding Agent, and Critic Model, which are optimized through K-Times Sampling, Perplexity as Rewards, DPO Training, Action Refinement, and Supervised Fine-tuning.
- The framework alternates between Planning Agent Alignment, which optimizes the planning agent to generate better subgoal sequences, and Grounding Agent Improving, which enhances the grounding agent's generalization capability using diverse, critic-corrected subgoal-action pairs.
- This joint alignment tuning process ensures a non-decreasing and progressively convergent training, leading to improved coordination and holistic performance across various tasks.
TrEnv: Transparently Share Serverless Execution Environments Across Different Functions and Nodes
- TRENv (Transparently Share Serverless Execution Environments): introduces a serverless platform that transparently shares execution environments across different functions and nodes, leveraging repurposable sandboxes, an mm-template API, CXL/RDMA memory pools, OS/hypervisor enhancements, CRIU integration, rootfs reconfiguration, cgroup optimization, browser sharing, and virtio-pmem devices to minimize startup latency and memory overhead for LLM agents.
- The platform optimizes for both container- and VM-based environments by enabling fast reuse and restoration of execution environments through memory templates and repurposable sandboxes, significantly reducing P99 latency and memory usage.
- TRENv addresses the high overhead of serverless computing for LLM agents by tackling cold starts, memory stranding, and state duplication, making serverless deployments more cost-efficient and scalable.
Combating the Memory Walls: Optimization Pathways for Long-Context Agentic LLM Inference
- PLENA (Programmable Long-context Efficient Neural Accelerator): introduces a hardware-software co-designed system that addresses memory walls in long-context agentic LLM inference through a flattened systolic array, an asymmetric quantization scheme, and native FlashAttention support, supported by a full toolchain including a custom ISA, compiler, simulator, and DSE flow.
- The system achieves high utilization by tailoring its flattened systolic array architecture to "fat GEMMs" and employs an asymmetric quantization scheme with mixed data types to reduce memory bandwidth and capacity limitations.
- PLENA's comprehensive toolchain, including a custom ISA and compiler, enables rapid adaptation and optimization for emerging Transformer models, delivering significantly higher throughput and utilization than existing accelerators.
MetaRAG: Metamorphic Testing for Hallucination Detection in RAG Systems
- MetaRAG (Metamorphic Testing for Hallucination Detection in Retrieval-Augmented Generation Systems): introduces a real-time, unsupervised, black-box metamorphic testing framework for hallucination detection in RAG systems, which includes Factoid Extraction, Mutation Generation, Factoid Verification, Score Calculation, and Identity-Aware Safeguards.
- The framework decomposes RAG answers into atomic factoids, generates controlled mutations (synonym/antonym substitutions), verifies each variant against the retrieved context, and aggregates inconsistencies into a response-level hallucination score.
- MetaRAG's span-level detection enables identity-aware safeguards by localizing unsupported claims and translating scores into topic-conditioned deployment policies, such as stricter thresholds or forced citations.
- TAM-Bench (Adaptive ML Benchmarks): introduces a diverse, realistic, and structured benchmark for evaluating LLM-based agents on end-to-end ML tasks, with components including Benchmark Source, Automated Web Scraping (featuring an LLM-based Agent Layer), Raw Content, Task Standardization (with LLM Transform), AutoML Agent, and Evaluator (incorporating LLM-as-a-Judge Constraint Pass), designed to address limitations in existing benchmarks.
- The framework leverages browser automation and LLMs for automated task collection and standardization, leaderboard-based difficulty modeling, and a multi-dimensional evaluation framework to assess agent capabilities holistically.
- TAM-Bench constructs benchmark subsets (Lite, Medium, Full) from 150 curated AutoML tasks, ensuring balanced coverage across data modalities and difficulty levels for robust and scalable evaluation.
Can Multimodal LLMs See Materials Clearly? A Multimodal Benchmark on Materials Characterization
- MatCha: introduces a multimodal benchmark for materials characterization image understanding, comprising task construction (defining tasks, extracting terms), data curation (collecting and processing HTML, figures, captions, supplementary datasets), and question generation (using GPT-4o, AI filtering, and expert review).
- The benchmark features 1,500 expert-level multiple-choice questions across 21 distinct tasks, reflecting real-world scientific challenges in materials research, covering processing correlation, morphology, structure, and property analysis.
- Evaluations on MatCha reveal a significant performance gap between state-of-the-art MLLMs and human experts, particularly in tasks requiring higher-level expertise and sophisticated visual perception, highlighting limitations in domain knowledge and reasoning.
LightAgent: Production-level Open-source Agentic AI Framework
- LightAgent (Production-level Open-source Agentic AI Framework): introduces a lightweight yet powerful agentic framework, integrating Agent, Memory (mem0), Tools, Tool Generator, Tree of Thought (ToT), LightSwarm, and LLMs to streamline multi-agent application development by resolving the trade-off between flexibility and simplicity.
- The framework redefines efficiency through a minimalist architecture, enabling autonomous tool generation, multi-agent collaboration, and robust fault tolerance with a 100% Python codebase of only 1,000 lines.
- LightAgent ensures rapid deployment across diverse scenarios by supporting dynamic agent specialization, multi-modal data handling, and compatibility with major LLMs and streaming APIs.
Harnessing Uncertainty: Entropy-Modulated Policy Gradients for Long-Horizon LLM Agents
- EMPG (Entropy-Modulated Policy Gradients): introduces a framework that re-calibrates the learning signal for LLM agents based on step-wise uncertainty and the final task outcome, addressing the coupling between gradient magnitude and policy entropy.
- This framework includes Self-Calibrating Gradient Scaling to dynamically adjust policy updates and a Future Clarity Bonus to guide agents towards predictable solution paths.
- EMPG's approach amplifies updates for confident correct actions, penalizes confident errors, and attenuates updates from uncertain steps, leading to more efficient and stable learning in long-horizon tasks.
- JUPITER: introduces a framework that enhances LLM data analysis capabilities by formulating data analysis as a search problem, leveraging Monte Carlo Tree Search for trajectory collection and a Value Model for inference-time value-guided search.
- The framework utilizes a fine-tuned LLM to generate thought-action pairs within a Jupyter Context, which are then executed in a Code Execution Environment to explore solution paths.
- JUPITER's Value Model, trained on MCTS-generated trajectories and Q-values from the NbQA Dataset, efficiently guides the search process, enabling open-source LLMs to achieve competitive performance with commercial agent systems on complex multi-step data analysis tasks.
Agentic LLMs for Question Answering over Tabular Data
- Agentic NL-to-SQL Pipeline: introduces a multi-stage LLM-driven framework for Question Answering over Tabular Data, including example selection, SQL query generation, answer extraction and formatting, answer verification, and answer reprocessing.
- This pipeline leverages LLMs (GPT-40, GPT-40-mini, DeepSeek v2:16b) for dynamic SQL query generation and answer refinement, utilizing embedding-based similarity for context and Chain-of-Thought prompting for enhanced reasoning.
- The system integrates a verification mechanism and iterative reprocessing to ensure query correctness, improve robustness across diverse table structures, and maximize answer accuracy.
Enabling Regulatory Multi-Agent Collaboration: Architecture, Challenges, and Solutions
- BEMAR (Blockchain-Empowered Multi-Agent Regulation): introduces a blockchain-enabled layered architecture for regulatory agent collaboration, comprising an Agent Layer (manages agents and collects data), a Blockchain Data Layer (maintains an immutable ledger), and a Regulatory Application Layer (provides advanced functionalities).
- The framework integrates three key modules: an agent behavior tracing and arbitration module for automated accountability, a dynamic reputation evaluation module for trust assessment, and a malicious behavior forecasting module for early adversarial detection.
- This architecture establishes a systematic foundation for trustworthy, resilient, and scalable regulatory mechanisms in large-scale LLM-empowered agent ecosystems by leveraging blockchain's transparency and immutability.
- LLM-SBP Methodology (Large Language Model - Scenario-Based Programming Methodology): introduces a hybrid approach for software development that integrates LLMs with Scenario-Based Programming, leveraging LLMs for code generation and strategic reasoning while mitigating their weaknesses through structured, modular, and verifiable design.
- The methodology involves a human-guided, modular workflow with components like structured prompts, iterative refinement, and human-in-the-loop feedback to reduce hallucinations and improve logical consistency.
- It enables developers to define high-level behavioral goals, decompose them into modular scenario objects, provide background knowledge to the LLM, and incrementally develop and refine scenario threads with continuous testing and formal verification.
AI Reasoning for Wireless Communications and Networking: A Survey and Perspectives
- AI Reasoning for Wireless Communications and Networking: introduces a comprehensive survey of reasoning-enabled AI in wireless communication networks, covering prompting strategies, architectural approaches, and learning paradigms, where the paper systematically categorizes and examines AI reasoning methods and their layered applications from the physical to the application layer.
- The survey highlights how LLM-based agents can combine reasoning with long-term planning, memory, tool utilization, and autonomous cross-layer control to dynamically optimize network operations with minimal human intervention.
- It addresses the limitations of traditional AI in dynamic environments, interpretability, and generalization, charting a path for integrating advanced reasoning techniques into next-generation wireless networks.
Strategic Tradeoffs Between Humans and AI in Multi-Agent Bargaining
- Multi-Agent Bargaining Game Evaluation Framework: introduces a novel multi-player bargaining game to directly compare human, LLM, and Bayesian agent performance and behavioral dynamics in dynamic negotiation settings.
- The framework evaluates agents based on surplus trajectories, trading patterns, and regret minimization, revealing distinct strategic approaches despite similar aggregate outcomes for humans and LLMs.
- Bayesian agents achieve the highest surplus through aggressive optimization, while LLMs favor conservative, concessionary trades, and humans employ more strategic, risk-taking, and fairness-oriented behaviors.
CDE: Curiosity-Driven Exploration for Efficient Reinforcement Learning in Large Language Models
- CDE (Curiosity-Driven Exploration): introduces a framework that enhances Reinforcement Learning with Verifiable Rewards (RLVR) in LLMs by leveraging intrinsic curiosity signals from both the Actor (using a Perplexity Bonus) and the Critic (using a Variance Bonus from a Multi-head Critic) to guide exploration.
- The framework integrates these curiosity signals as an exploration bonus within existing RL algorithms like GRPO and PPO, mitigating issues such as premature convergence and entropy collapse in LLM training.
- CDE's theoretical analysis demonstrates its ability to penalize overconfident errors and encourage diverse correct responses, leading to improved calibration and consistent performance gains on mathematical reasoning benchmarks.
- TMK (Trait Modulation Key) Framework: introduces a modular prompting framework that systematically controls LLM-based Conversational Agents' (CAs) personality across Big Five traits at low, medium, and high expression levels using Personality Keys and Style Cues Keys.
- The framework enables nuanced personality expression in LLMs, moving beyond binary trait manipulations to investigate the impact of personality expression and user-agent alignment on user perceptions in goal-oriented tasks.
- The study utilizes a text-only chat interface for user interaction with CAs, evaluating user perceptions across measures like Intelligence, Enjoyment, and Trust, and assessing personality alignment via Euclidean distance.
LLMs as Agentic Cooperative Players in Multiplayer UNO
- LLM UNO: introduces a framework that enables decoder-only LLMs to act as agentic cooperative players in the RLCard UNO environment, receiving game state and selecting actions via text prompts.
- The framework evaluates LLMs ranging from 1B to 70B parameters in both autonomous play against a random agent and cooperative play assisting a rule-based teammate in a three-player game.
- It investigates how model scale and prompting techniques, specifically cloze and counterfactual prompting, influence LLM performance and cooperative effectiveness in strategic decision-making.
Latency and Token-Aware Test-Time Compute
- Latency and Token-Aware Inference-Time Scaling Framework: introduces a latency- and token-aware approach for inference-time scaling, dynamically allocating compute and selecting methods per query using its Query Input, Decoding Strategies, Utility Predictors, Utility Function, and Optimal Strategy Selector, supported by an LLM, PRM, and Embedding Backbones.
- The framework explicitly incorporates both token cost and wall-clock latency into its utility formulation, which is crucial for user experience and efficient agentic workflows requiring multiple LLM queries.
- Experiments on reasoning benchmarks demonstrate that this query-adaptive approach consistently outperforms static strategies, achieving favorable accuracy-cost trade-offs.
SWE-Effi: Re-Evaluating Software AI Agent System Effectiveness Under Resource Constraints
- SWE-Effi introduces a multi-dimensional framework for re-evaluating AI systems in software engineering, incorporating Core Performance Metrics (raw measurements) and Resource Effectiveness Metrics (derived efficiency scores) to assess holistic effectiveness.
- This framework defines effectiveness as the balance between solution accuracy (e.g., resolve rate) and consumed resources (e.g., tokens, time, cost), addressing the limitations of traditional single-metric evaluations.
- The paper applies SWE-Effi to re-rank popular AI systems on a SWE-bench subset, revealing insights into LLM-scaffold synergy, the "Token Snowball" effect, and "expensive failures" where unresolved tasks consume excessive resources.
Meta-Learning Reinforcement Learning for Crypto-Return Prediction
- Meta-RL-Crypto: introduces a unified transformer-based architecture that unifies meta-learning and reinforcement learning to create a self-improving trading agent, featuring an Actor, Judge, and Meta-Judge in a closed-loop system.
- This framework leverages multimodal market inputs and internal preference feedback, continuously refining its trading policy and evaluation criteria without human supervision.
- A multi-objective reward design, incorporating profitability, risk control, liquidity, and sentiment alignment, prevents reward hacking and promotes robust trading behavior.
Musculoskeletal simulation of limb movement biomechanics in Drosophila melanogaster
- Drosophila Leg Musculoskeletal Modeling Pipeline: introduces a 3D, data-driven musculoskeletal model of Drosophila legs, implemented in OpenSim and MuJoCo, which incorporates Hill-type muscle representation and is optimized using morphological imaging and 3D pose estimation data.
- This model enables muscle-actuated behavioral replay, predicts coordinated muscle synergies across diverse behaviors, and investigates the impact of passive joint properties on learning speed.
- The framework integrates anatomical, physiological, and behavioral data into a unified modeling approach, providing insights into motor control biomechanics and facilitating the control of embodied artificial agents.
- SART (Self-Augmented Robot Trajectory): introduces a framework where a Human Expert provides a Single Human Demonstration and Precision Boundaries to enable a Robot Self-Augmentation Module to generate an Augmented Training Dataset for learning a robust Parametric Policy on a Robot System.
- This framework minimizes human effort by requiring only one demonstration and subsequent annotation via an Interactive Annotation Interface, followed by autonomous data expansion within defined safe regions.
- SART demonstrates higher success rates in clearance-limited robotic manipulation tasks across both Simulation Environment and Real-world Environment, improving data collection efficiency and policy robustness.
Curriculum-Based Multi-Tier Semantic Exploration via Deep Reinforcement Learning
- VLM-DRL framework: introduces a novel Deep Reinforcement Learning architecture for resource-efficient semantic exploration, integrating a Vision-Language Model (VLM) common-sense through a layered reward function and a curriculum learning strategy.
- The framework enables an agent to strategically query the VLM for external guidance, optimizing resource usage, and progressively develops exploration skills through distinct training phases.
- This approach enhances object discovery rates and guides the agent towards semantically rich regions, demonstrating a scalable method for embedding common-sense reasoning in autonomous exploration.
Flip Co-op: Cooperative Takeovers in Shared Autonomy
- Flip Co-op (Cooperative Takeovers in Shared Autonomy): introduces a game-theoretic framework for modeling cooperative takeovers in shared autonomy, including dynamic game formulation, human agent, autonomous agent, FlipDyn state, takeover actions, stochastic human intent, Nash equilibrium (NE) strategies, saddle-point value functions, linear-quadratic (LQ) system model, bimatrix potential game reformulation, and unifying potential function, where it formulates control switching as a dynamic game to derive Nash equilibrium-based takeover strategies.
- The framework establishes the existence and characterization of NE under stochastic human intent and provides closed-form recursions for LQ systems, enabling efficient computation of cooperative takeover policies.
- It further extends the model to partially misaligned utilities through a bimatrix potential game reformulation, demonstrating its applicability to a vehicle trajectory tracking problem.
ProgD: Progressive Multi-scale Decoding with Dynamic Graphs for Joint Multi-agent Motion Forecasting
- ProgD (Progressive Multi-scale Decoding with Dynamic Graphs for Joint Multi-agent Motion Forecasting): introduces a novel progressive multi-scale decoding strategy for joint multi-agent motion forecasting, utilizing dynamic heterogeneous graphs to model evolving social interactions and progressively eliminate uncertainty.
- The framework employs an encoder-decoder architecture, where the encoder processes historical agent states and road networks, and the decoder forecasts future motions through a temporal module, dynamic graph construction, and dual-stage graph convolution modules.
- ProgD achieves state-of-the-art performance by adaptively improving the modeling of evolving context and enhancing prediction accuracy and consistency through its multi-scale decoding procedure.
Occupancy-aware Trajectory Planning for Autonomous Valet Parking in Uncertain Dynamic Environments
- Occupancy-aware Trajectory Planning Framework: introduces a system for autonomous valet parking in dynamic, uncertain environments, with Static Map (prior knowledge of parking lot), Prediction Model (dynamic agent states and predictions), Spot Occupancy Estimator (recursive belief estimation of spot occupancy), Strategy Planner (action selection based on occupancy forecasts), Path Planner (trajectory generation), and Reference Tracker (vehicle control).
- The framework predicts future parking spot occupancy by distinguishing between initially vacant and occupied spots, leveraging predicted motion of dynamic agents, and incorporating partial, noisy observations within a limited Field-of-View model.
- It adaptively balances goal-directed parking maneuvers with exploratory navigation based on information gain, intelligently incorporating wait-and-go behaviors at promising spots to improve parking efficiency and safety.
Mind Meets Space: Rethinking Agentic Spatial Intelligence from a Neuroscience-inspired Perspective
- Agentic Spatial Intelligence Framework: introduces a novel computational framework grounded in neuroscience principles, designed to advance agentic spatial intelligence for better interaction with the physical 3D world, comprising bio-inspired multimodal sensing, multi-sensory integration, egocentric-allocentric conversion, an artificial cognitive map, spatial memory, and spatial reasoning.
- This framework mimics human spatial cognition by processing diverse sensory inputs, transforming raw data into structured representations, converting first-person views into world-centered maps, building an internal spatial model, maintaining adaptive long-term memory, and enabling goal-oriented decision-making.
- The framework provides a structured pathway for developing human-like spatial intelligence in AI agents, addressing limitations of current systems in representing spatial structures, reasoning about spatial relationships, and planning in spatial contexts.
- Agentic Annotation Evaluation Paradigm: introduces a novel method for evaluating LLM annotation quality without oracle feedback, utilizing a Noisy Teacher (LLM) (generates noisy annotations), a Student Model (MINILM) (evaluates, assigns annotations), a User Preference Distribution (limited labeled data), an Average Similarity (AS) Function (calculates similarity for voting), a Consistent and Inconsistent (CAI) Ratio (measures annotation reliability), a Group Prompting Mechanism (teacher LLM annotation strategy), Consistent Samples (teacher-student agreement), and Inconsistent Samples (teacher-student disagreement).
- The Student Model acts as an unsupervised feedback mechanism, employing a user preference-based majority voting strategy to assess the consistency of the LLM's outputs.
- The CAI Ratio quantifies annotation quality and serves as a critical tool for model selection, demonstrating a strong positive correlation with LLM accuracy in unsupervised settings.
- AgentGym-RL: introduces a unified, modular, and flexible end-to-end RL framework for training LLM agents for multi-turn interactive decision-making, encompassing an LLM Agent, an Environment Module, and a Training Module.
- The framework supports diverse real-world scenarios like web navigation, deep search, digital games, embodied tasks, and scientific tasks, and integrates mainstream RL algorithms such as PPO, GRPO, RLOO, and REINFORCE++.
- It also proposes ScalingInter-RL, a progressive interaction-scaling approach that adaptively adjusts the agent-environment interaction horizon to balance exploration-exploitation and enhance optimization stability.
ChemBOMAS: Accelerated BO in Chemistry with LLM-Enhanced Multi-Agent System
- ChemBOMAS (LLM-Enhanced Multi-Agent System for accelerating BO in chemistry): introduces a novel framework that synergistically integrates knowledge-driven coarse-grained optimization and data-driven fine-grained optimization, where LLMs intelligently decompose the search space and generate pseudo-data points to accelerate Bayesian Optimization in chemistry.
- The framework employs a multi-agent system where LLM-powered agents reason over chemical knowledge to identify promising candidate regions and enhance the BO process by generating informative pseudo-data points.
- This approach significantly improves data utilization efficiency and accelerates convergence, validated through benchmark evaluations and wet-lab experiments on challenging chemical reactions.
SWE-Mirror: Scaling Issue-Resolving Datasets by Mirroring Issues Across Repositories
- SWE-MIRROR: introduces a three-phase pipeline including Task Collection, Task Mirroring, and Task Validation, which distills real-world issues from GitHub, mirrors them into configured Gym environments using LLMs and agents, and re-animates them as verifiable issue-resolving tasks.
- The framework leverages LLMs like Qwen3-32B, GPT-40-2024-0513, and GPT-4.1, along with specialized Test and Mirror Agents, to generate key artifacts such as
test.patch,mirror.patch,fix.patch, and aproblem_statementfor each mirrored task. - The approach breaks the one-to-one dependency between task context and Gym environments, enabling the creation of a large-scale, verifiable dataset (SWE-MIRROR-60K) that significantly improves LLM-based coding agents' issue-resolving capabilities and cross-lingual generalization.
Architecting Resilient LLM Agents: A Guide to Secure Plan-then-Execute Implementations
- P-t-E (Plan-then-Execute): introduces a secure architectural pattern for LLM agents, separating strategic planning by a Planner from tactical execution by an Executor, complemented by a Verifier, Refiner, and Re-planner for resilience and security.
- The pattern integrates robust security controls like Input Sanitization and Validation, Output Filtering, Dual LLM Pattern, Principle of Least Privilege, Task-scoped Tool Access, Role-Based Access Control (RBAC), Sandboxed Execution via Docker Containers, and Human-in-the-Loop (HITL) verification.
- P-t-E supports advanced patterns such as Directed Acyclic Graphs (DAGs) for parallel execution, GraphQL Integration for optimized data retrieval, and Graph-Based Conditional Execution Paths, with implementations detailed across LangGraph, CrewAI, and AutoGen frameworks.
AutoODD: Agentic Audits via Bayesian Red Teaming in Black-Box Models
- AutoODD: introduces an LLM-Agent centric framework for automated generation of semantically relevant test cases to search for failure modes in specialized black-box models, leveraging an Agent (orchestrates testing) that interacts with a Memory (stores past interactions), a Failure Tracker (monitors failure occurrences), an Encoder (embeds text prompts), UMAP Reduction (reduces embedding dimensions), GP Fitting (models failure landscape), a Prompt2Input (generates SUT inputs), and a Testing (evaluates SUT performance) module to audit a SUT (black-box model) and record Results (SUT pass/fail outcomes) from generated Sample #1 (generated test case) and Sample #2 (generated test case).
- The framework combines LLM-Agent orchestration with Bayesian uncertainty estimation to efficiently explore the failure landscape of black-box models within semantically meaningful embedding spaces.
- This approach aims to discover operational boundaries and failure modes in safety-critical systems with significantly reduced sample complexity, providing a scalable methodology for verifying model reliability.
- Agents of Discovery: introduces an agentic framework where a team of LLM agents collaboratively solves data analysis-based research problems in high-energy physics, utilizing specialized tools and a local execution environment.
- The framework includes a Researcher orchestrating tasks, a Coder writing Python code, and Code and Logic Reviewers providing iterative feedback to refine analysis strategies and ensure reproducibility.
- Evaluated on the LHC Olympics anomaly detection challenge, the system demonstrates the capacity of LLMs to automate complex scientific workflows, with advanced models achieving human state-of-the-art performance.
HUMANAGENCYBENCH: Scalable Evaluation of Human Agency Support in AI Assistants
- HAB (HUMANAGENCYBENCH): introduces a scalable and adaptive benchmark for evaluating human agency support in AI assistants, including a Simulator LLM (generates user query test candidates), Validation Rubric (criteria for test candidate quality), Validator LLM (scores and filters test candidates), Diversity Sampling (k-means clustering) (selects representative queries), Dimension Test Set (500 simulated user queries per dimension), and an Evaluation Model (LLM) (scores LLM-based assistant responses) across six dimensions of human agency.
- HAB operationalizes human agency into six key dimensions: Ask Clarifying Questions, Avoid Value Manipulation, Correct Misinformation, Defer Important Decisions, Encourage Learning, and Maintain Social Boundaries, each with specific evaluation rubrics to assess LLM-based assistants.
- The framework leverages LLMs for automated test generation and evaluation, providing a systematic approach to assess how different AI assistants support or diminish user control and autonomy in various interaction scenarios.
Co-Investigator AI: The Rise of Agentic AI for Smarter, Trustworthy AML Compliance Narratives
- Co-Investigator AI: introduces an agentic framework for generating Suspicious Activity Reports (SARs), with a Data Ingestion & Structuring Layer (ingests, transforms raw data), AI-Privacy Guard Layer (identifies, anonymizes sensitive data), Crime Type Detection Agent (extracts risk indicators, classifies typologies), Planning Agent (orchestrates agents, allocates resources), Specialized Typology Detection Agents (analyze specific crime types), External Intelligence Agent (accesses external risk intelligence), Narrative Generation Agent (synthesizes SAR drafts), Compliance Validation Agent (Agent-as-a-Judge) (verifies narrative quality, compliance), Feedback Agent (integrates human feedback), Dynamic Memory Management (maintains regulatory, historical, typology memory), Analytical Tools (extracts risk indicators, searches intelligence, links accounts), and User Interface (human investigator interaction), designed to produce SARs faster and with greater accuracy while maintaining human interpretability.
- The framework leverages a modular, human-in-the-loop design, integrating specialized AI agents for planning, crime type detection, external intelligence gathering, and compliance validation, supported by dynamic memory and an AI-Privacy Guard layer for sensitive data handling.
- This approach aims to streamline SAR drafting, align narratives with regulatory expectations, and enable compliance teams to focus on higher-order analytical work, marking a shift towards scalable, reliable, and transparent SAR generation in Anti-Money Laundering (AML) compliance.
A Systematic Survey on Large Language Models for Evolutionary Optimization: From Modeling to Solving
- Systematic Survey Taxonomy: introduces a comprehensive review of LLMs for evolutionary optimization, categorizing research into LLMs for Optimization Modeling (automatically transforming natural language to mathematical models) and LLMs for Optimization Solving (enhancing or directly performing optimization tasks).
- The LLMs for Optimization Solving category is further subdivided into LLMs as Optimizers (solving problems via iterative natural language interaction), Low-level LLM-assisted Optimization Algorithms (embedding LLMs for specific operations), and High-level LLM-assisted Optimization Algorithms (orchestrating algorithm selection and generation).
- This taxonomy provides a structured framework for understanding the evolving landscape of LLM applications in optimization, highlighting current challenges and future directions towards self-evolving agentic ecosystems.
Exploratory Retrieval-Augmented Planning For Continual Embodied Instruction Following
- ExRAP (Exploratory Retrieval-Augmented Planning): introduces an embodied agent framework for continual instruction following in dynamic environments, utilizing a Temporal Embodied Knowledge Graph (TEKG) for environmental context memory and an exploration-integrated task planning scheme.
- The framework enhances LLMs' embodied reasoning by decomposing continual instructions into queries and executions, which are evaluated by a Memory-augmented query evaluator using the TEKG and refined with temporal consistency.
- It integrates an Exploitation planner and an Exploration planner to balance task achievement and environmental knowledge acquisition, demonstrating superior performance in goal success and execution efficiency.
Componentization: Decomposing Monolithic LLM Responses into Manipulable Semantic Units
- Componentization: introduces an approach that decomposes monolithic LLM responses into modular, independently editable units, leveraging MAOD (Modular and Adaptable Output Decomposition) for semantic segmentation and CBRA (Component-Based Response Architecture) for structured workflow management.
- The MAODchat reference prototype implements CBRA using a microservices architecture, featuring a Flask-based Frontend, FastAPI Backend, FastAPI MAOD Agent, PostgreSQL Database, and a Caddy Reverse Proxy for orchestration.
- This framework enhances human-AI collaboration by enabling user-driven component manipulation (edit, select/toggle, regenerate) and dynamic recomposition, fostering content and architectural resilience against "catastrophic regeneration" failures.
Global Constraint LLM Agents for Text-to-Model Translation
- Global Constraint LLM Agents: introduces an agentic framework for translating natural language descriptions of optimization problems into MiniZinc models, utilizing specialized LLM agents for global constraint detection and code generation, and an assembler LLM for model integration.
- This framework decomposes the complex text-to-model translation task into smaller, manageable sub-tasks, where each specialized LLM agent focuses on a specific global constraint type, simplifying reasoning.
- The approach demonstrates improved performance over baseline prompting strategies, including chain-of-thought, by reducing cognitive load on individual LLMs and enabling a collaborative modeling process.
- GeoJSON Agents (A Multi-Agent LLM Architecture for Geospatial Analysis): introduces an automated spatial analysis framework, with Planner and Worker agents, that transforms natural language tasks into structured GeoJSON operation commands and processes spatial data using Function Calling or Code Generation.
- The framework leverages a Planner agent for task decomposition and a Worker agent for execution, employing either predefined function APIs via Function Calling or dynamically generated Python code via Code Generation.
- This multi-agent LLM architecture significantly outperforms general-purpose LLMs in geospatial analysis, offering enhanced performance and scalability for GIS automation by integrating GeoJSON data.
- HYPOGENEAGENT (Hypothesis Language Agent): introduces a multi-stage LLM-driven framework for gene-set cluster resolution selection, including an Input Gene List (input data for analysis), a General LLM agent (generates initial biological process explanations), a Hypothesis LLM agent (refines and ranks top GO hypotheses with confidence scores), and Output LLM-proposed descriptions (final ranked biological process descriptions).
- The framework quantifies biological relevance by combining semantic similarity, intra-cluster agreement, and inter-cluster distinctiveness to select optimal cluster resolutions and provide biologically informed interpretations.
- HYPOGENEAGENT automates the annotation process, bridging the gap between unsupervised partitioning and biologically informed interpretation in single-cell multi-omics studies.
A Role-Aware Multi-Agent Framework for Financial Education Question Answering with LLMs
- Role-Aware Agent Framework: introduces a multi-agent system for financial education QA, including an Evidence Retriever (retrieves topic-relevant evidence), a Base Generator (drafts initial answer), and an Expert Reviewer (critiques and refines answer).
- This framework leverages role-based prompting to assign domain-specific personas to LLMs, enhancing reasoning and factual accuracy in financial problem-solving.
- The system demonstrates improved answer accuracy and explanation quality by integrating retrieved evidence and expert critique, outperforming zero-shot Chain-of-Thought baselines.
Dual-Stage Safe Herding Framework for Adversarial Attacker in Dynamic Environment
- Dual-Stage Safe Herding Framework: introduces a hierarchical hybrid strategy framework for cooperative capture and escort, enabling safe guidance of adversarial agents to designated regions in dynamic, obstacle-dense environments.
- The framework decomposes the problem into a capture stage for initial encirclement and an escort stage for continuous guidance, integrating reach-avoid game theory and local motion planning.
- It utilizes a virtual containment boundary and event-triggered pursuit mechanisms for scalable and robust multi-agent coordination against attackers with unknown evasion strategies.
Leveraging AI Agents for Autonomous Networks: A Reference Architecture and Empirical Studies
- AN Agent reference architecture: introduces a novel dual-driver architecture for Autonomous Networks, deploying coordinated proactive-reactive runtimes driven by hybrid knowledge representation.
- The framework is empirically validated through a Radio Access Network (RAN) Link Adaptation (LA) Agent, demonstrating sub-10 ms real-time control and significant performance enhancements in 5G NR.
- This architecture integrates self-awareness, choice-making, and decision-making mechanisms, leveraging LLMs and various tools to achieve L4 operational autonomy and overcome traditional automation barriers.
A Comprehensive Review of Reinforcement Learning for Autonomous Driving in the CARLA Simulator
- A Comprehensive Review of Reinforcement Learning for Autonomous Driving in the CARLA Simulator: introduces a systematic analysis of around 100 peer-reviewed papers, categorizing RL literature by algorithmic family, state/action/reward formulations, evaluation metrics, and CARLA scenarios.
- The review highlights that over 80% of studies rely on model-free RL methods and details the diverse sensor modalities, control abstractions, and reward shaping techniques used in CARLA-based autonomous driving research.
- It consolidates the evaluation landscape, distills persistent challenges like sparse rewards and sim-to-real transfer, and outlines promising future directions for advancing RL-based autonomous driving towards real-world deployment.
AgentSentinel: An End-to-End and Real-Time Security Defense Framework for Computer-Use Agents
- AgentSentinel: introduces an end-to-end, real-time security defense framework for computer-use agents, employing a client-server architecture with a Monitor, Instrumented Agent, Tracer, and Auditor to intercept sensitive operations and conduct comprehensive security audits.
- The framework's auditing mechanism correlates current task context with system traces, integrating rule-based and LLM-based approaches, further optimized by a security query cache and QPS optimizer for real-time threat detection.
- AgentSentinel demonstrates an average defense success rate of 79.6% against diverse attack scenarios, significantly outperforming existing baseline defense mechanisms for LLM-based computer-use agents.
VeriOS: Query-Driven Proactive Human-Agent-GUI Interaction for Trustworthy OS Agents
- VeriOS-Agent: introduces a query-driven human-agent-GUI interaction framework that enables OS agents to determine when to query humans for more trustworthy task completion, built upon a Two-Stage Learning Paradigm that decouples and utilizes Scenario Knowledge and Action Knowledge from the VeriOS-Bench dataset to train an Automated OS Agent.
- The framework allows the VeriOS-Agent to autonomously execute actions in normal conditions while proactively querying humans in untrustworthy scenarios, leveraging human responses to ensure reliable task completion via the GUI.
- This approach improves step-wise success rates in untrustworthy scenarios without compromising normal performance, demonstrating strong generalization and scalability for trustworthy OS agent operation.
Astra: A Multi-Agent System for GPU Kernel Performance Optimization
- Astra (A Multi-Agent System for GPU Kernel Performance Optimization): introduces, "Astra (A Multi-Agent System for GPU Kernel Performance Optimization)", with Testing Agent (creates test cases), Profiling Agent (measures performance), Planning Agent (proposes modifications), and Coding Agent (generates new kernels), where Astra optimizes existing CUDA GPU kernels from SGLang through iterative code generation, testing, profiling, and planning.
- This multi-agent system achieves an average speedup of 1.32x using zero-shot prompting with OpenAI o4-mini, demonstrating autonomous application of loop transformations, memory access pattern optimization, CUDA intrinsics, and fast math operations.
- The framework focuses on optimizing existing CUDA implementations rather than generating them from PyTorch modules, addressing a critical challenge in LLM serving and training efficiency.
- DREAMS (Decentralized Resource Allocation and Service Management): introduces a decentralized framework for optimizing microservice placement across the compute continuum, featuring Local Domain Managers, Administrative, Configuration Control, Observability and Diagnostics, Domain Monitoring, Migration Intelligence, Consensus Management, Migration Execution, Inter-Domain Communication, and Recovery and Fault Tolerance Modules, along with various repositories.
- The framework enables autonomous agents (LDMs) within each computational domain to collaboratively make service placement decisions using a Raft-based consensus algorithm and cost-benefit voting, ensuring responsive, privacy-preserving, and fault-tolerant coordination.
- DREAMS achieves globally optimized service placements while maintaining high fault tolerance and sub-linear scalability for key coordination operations like LDM registration and migration voting, making it suitable for multi-stakeholder, dynamic manufacturing environments.
Autonomous Code Evolution Meets NP-Completeness
- SATLUTION (Autonomous Code Evolution Framework): introduces a repository-scale, self-evolving coding framework via LLMs, with Planning and Coding Agent, Planning Claude Model, Coding Claude Model, Cursor Environment, Self-evolved Rulebase, Static Initialization Rules, Dynamic Self-Evolved Rules, SATLUTION Repository, Two-stage Verification Pipeline, Compilation Check, Smoke Test, Full Correctness Validation, SAT Assignment Verifier, DRAT Proof Checker, Distributed Runtime Evaluator, Feedback Metrics, Post-evaluation Analyzer, Rule Update Engine, and Rule Version Manager, designed to autonomously evolve SAT solver repositories under strict correctness guarantees and distributed runtime feedback.
- The framework orchestrates LLM agents through Planning and Coding stages, guided by a self-evolving rule system and a two-stage verification pipeline, to iteratively improve SAT solver engineering at a full repository scale.
- SATLUTION successfully evolved SAT solvers that outperformed human-designed winners of the SAT Competition 2025, demonstrating the potential of AI agents for champion-level performance in NP-complete problem solving.
Guided Reasoning in LLM-Driven Penetration Testing Using Structured Attack Trees
- STT-based reasoning pipeline: introduces a guided reasoning pipeline for LLM-driven penetration testing, incorporating a deterministic Structured Task Tree (STT) to constrain the LLM's reasoning process to explicitly defined tactics, techniques, and procedures from the MITRE ATT&CK Matrix.
- This framework enhances the accuracy and efficiency of automated cybersecurity assessments by guiding the LLM agent through a predefined task flow, reducing hallucinations and unproductive actions compared to self-guided reasoning methods.
- The pipeline enables smaller LLMs to perform complex, multi-step reasoning effectively and consistently, demonstrating significant improvements in subtask completion rates and requiring fewer model queries across various HackTheBox challenges.
KLIPA: A Knowledge Graph and LLM-Driven QA Framework for IP Analysis
- KLIPA (Knowledge Graph and LLM-Driven Question-Answering Framework for IP Analysis): introduces a novel framework for patent analysis, integrating a Patent Knowledge Graph (structured representation of patent data), a Retrieval-Augmented Generation (RAG) System (retrieves semantically relevant patent information), and a ReAct Agent Framework (dynamically determines retrieval strategy and generates responses) to enhance relationship identification, patent retrieval, and knowledge discovery.
- The framework's Patent Knowledge Graph (KG) is constructed from IP Dataset (raw input patent documents) via NER and RE (Named Entity Recognition and Relation Extraction) and stored in a Neo4j Database, while the LLM (Large Language Model) powers reasoning and response synthesis, supported by modules like Document Parser (processes various document formats) and Text Splitting Module (segments documents into chunks).
- The LLM-based QA agent further leverages a Vector Database (stores document embeddings for RAG) and an Embedding Generation Module (creates dense vector representations) for its Hybrid Retriever (combines vector similarity and keyword matching), with a Gradio-based User Interface (for interactive query handling) facilitating user interaction.
- LLM ATLAS (LLM Agency Theory-Led Alignment Strategy): introduces a conceptual framework grounded in agency theory to mitigate LLM alignment problems during organizational LLM adoption, by combining organizational LLM adoption phases and agency theory concepts to derive an LLM alignment problem-solution space.
- The framework identifies information asymmetries between the adopting organization (principal) and the black-box LLM (agent) as the root cause of alignment issues, categorizing them as hidden characteristics or hidden actions.
- It provides practical solutions like signaling through model cards and screening via adversarial attacks, or bonding through human preference incorporation and monitoring via model-driven supervision, tailored to specific LLM adoption phases.
AgentX: Towards Orchestrating Robust Agentic Workflow Patterns with FaaS-hosted MCP Services
- AgentX: introduces a novel agentic workflow pattern, composed of a Stage Generation Agent, a Planner Agent, and an Executor Agent, designed to orchestrate robust agentic workflows with FaaS-hosted Model Context Protocol (MCP) services.
- This framework decomposes user tasks into sequential stages, with agents collaboratively creating detailed plans and executing them using external tools, while actively managing context to prevent information overload and hallucinations.
- The paper evaluates AgentX's performance against state-of-the-art patterns, demonstrating its competitive or superior success rate, and explores FaaS deployment of MCP servers to enhance scalability, security, and accessibility for real-world applications.
TOWARDS GENERALIZED ROUTING: MODEL AND AGENT ORCHESTRATION FOR ADAPTIVE AND EFFICIENT INFERENCE
- MoMA (Mixture of Models and Agents): introduces a generalized routing framework that integrates both LLM and agent-based routing, effectively handling diverse queries through precise intent recognition and adaptive routing strategies.
- The framework employs a two-layer routing mechanism, first determining if an LLM can handle the query, then either selecting an optimal agent via a context-aware FSM or an optimal LLM based on a score-cost tradeoff.
- MoMA achieves an optimal balance between efficiency and cost by leveraging a detailed training dataset to profile LLM and agent capabilities, ensuring robust and scalable adaptive services.
FEED-O-METER: Fostering Design Feedback Skills through Role-playing Interactions with AI Mentee
- FEED-O-METER (Fostering Design Feedback Skills through Role-playing Interactions with AI Mentee): introduces a novel system that employs carefully designed LLM-based agents to create an environment for students to practice giving design feedback, featuring a User Interface (displays design information, enables chat, visualizes feedback), an LLM-based Pipeline (generates responses, categorizes feedback, extracts knowledge, updates ideas), an AI Mentee (LLM-based agent, novice design student persona), a Knowledge State (stores mentee's design expertise), and an Action Plan (tracks design refinement recommendations).
- The system allows users to role-play as mentors, providing feedback to an AI mentee and enabling them to reflect on how their feedback impacts the AI mentee's idea development process through real-time visualizations and counter-questions.
- By simulating a realistic, low-pressure feedback environment, FEED-O-METER aims to enhance students' design feedback skills, critical thinking, and self-reflection without the anxiety of real-world judgment.
Talking with Oompa Loompas: A novel framework for evaluating linguistic acquisition of LLM agents
- Oompa Loompas framework: introduces a novel experimental framework for evaluating the linguistic acquisition of LLM agents, featuring an LLM agent, a deterministic Oompa Loompa bot, the constructed Tinkatongue language, a feedback mechanism, a system prompt, a synthetic dataset, and evaluation metrics.
- The framework assesses an LLM agent's ability to learn Tinkatongue through pattern recognition and interactive feedback from the Oompa Loompa bot, which provides "koro" for valid sentences and "moko lira bani" for invalid ones.
- This evaluation method simulates human-like language acquisition, revealing that LLM agents adopt strategies like imitation and babbling, and highlighting the challenge of sustained language learning despite good feedback responsiveness.
Towards Post-mortem Data Management Principles for Generative AI
- Post-mortem Data Management Principles: introduces a framework for managing deceased individuals' data in Generative AI systems, encompassing an analysis phase, three core principles, and deployment strategies.
- The framework addresses data ownership, privacy, and ethical concerns for post-mortem data, proposing rights like data deletion and inheritance, alongside purpose limits.
- It recommends both regulatory enforcement and technical solutions, such as digital wills and privacy-preserving techniques, to implement these principles effectively.
- SpecifyUI (interactive system): introduces a vision-centered intermediate representation, SPEC, to make design intent explicit and controllable in UI generation, enabling users to extract specifications, compose elements into a coherent whole, and iteratively refine designs through direct selection and element extraction.
- The system leverages a multi-agent UI generation pipeline, including Region Segmentation, VLM, and LLM components, to translate structured specifications into high-fidelity UI designs, supporting targeted edits at global, regional, and component levels.
- SpecifyUI integrates a Retrieval-Augmented Generation (RAG) system and a Debug Agent to enhance generation fidelity and robustness by grounding LLM outputs with a SPEC-UI Code Database and self-correcting errors.
CancerGUIDE: Cancer Guideline Understanding via Internal Disagreement Estimation
- CancerGUIDE framework: introduces an LLM agent-based approach to automatically generate guideline-concordant treatment trajectories for non-small cell lung cancer patients, leveraging a meta-classifier to verify prediction accuracy with calibrated confidence scores.
- The framework addresses the evaluation bottleneck for LLM performance on guideline adherence by combining expensive human annotations with model consistency information, enabling scalable assessment without extensive expert annotation.
- It establishes a clinically viable framework for LLM-based guideline adherence systems that balance accuracy, interpretability, and regulatory requirements while reducing annotation costs and providing a scalable pathway toward automated clinical decision support.
- XML Prompting as Grammar-Constrained Interaction: introduces a logic-first treatment of XML prompting that unifies grammar-constrained decoding, fixed-point semantics over hierarchical prompts, and convergent human-AI interaction loops.
- The framework formalizes XML prompting as a typed tree language with a refinement order, defining prompt transformers (T) to capture interaction rounds and proving the existence of least fixed points for steady-state protocols.
- It further introduces a task-aware tree metric, demonstrating Banach-style convergence for iterative guidance and providing multi-layer human-AI interaction templates with correctness guarantees.
K2-Think: A Parameter-Efficient Reasoning System
- K2-THINK (K2-Think: A Parameter-Efficient Reasoning System): introduces a reasoning system built on a 32B parameter LLM, enhanced by Long Chain-of-thought Supervised Finetuning (SFT) (CoT training), Reinforcement Learning with Verifiable Rewards (RLVR) (reasoning performance enhancement), Plan-Before-You-Think (Agentic Planning) (structured prompt generation), Best-of-N Sampling (Test-time Scaling) (optimal response selection), Speculative Decoding (inference acceleration), and Cerebras Wafer-Scale Engine (WSE) (inference-optimized hardware).
- This system achieves frontier reasoning performance comparable to much larger models by synergistically combining advanced post-training and test-time computation techniques.
- K2-THINK prioritizes mathematical reasoning, achieving state-of-the-art scores on public benchmarks for open-source models while maintaining strong performance in code and science domains.
∆L Normalization: RETHINK Loss AGGREGATION IN RLVR
- ∆L Normalization: introduces a simple yet effective loss aggregation method tailored for Reinforcement Learning with Verifiable Rewards (RLVR) to address high gradient variance and unstable optimization caused by dynamic generation lengths in LLMs.
- The method provides an unbiased estimate of the true policy loss and minimizes gradient variance by applying specific normalization weights derived from response lengths and a tunable hyperparameter α.
- Extensive experiments demonstrate that the method consistently achieves superior results across different model sizes, maximum lengths, and tasks, promoting stable training and higher accuracy.
- DPH-RL (Diversity-Preserving Hybrid RL): introduces a novel framework that leverages mass-covering f-divergences (e.g., Forward-KL, JS-divergence) as a rehearsal mechanism, continuously referencing the initial policy to maintain broad solution coverage and mitigate diversity collapse in Reinforcement Learning with Verifiable Reward (RLVR).
- The framework operates in two stages: a pre-sampling stage that partitions a static dataset into Dpef (perfect examples) and Dexp (challenging examples), and an online training stage that applies distinct loss functions (Lexp for exploration and Lpef for knowledge retention) to these subsets.
- DPH-RL, implemented as DPH-F (Forward-KL) or DPH-JS (JS-divergence), improves multi-attempt performance (Pass@k) and single-attempt accuracy (Pass@1) both in- and out-of-domain, while being training-efficient by computing f-divergence using generator functions from the initial policy.
- DWM and DTA: introduces a novel distributed coordination method for multi-agent systems, addressing task assignments in dynamic environments with asymmetric obstacles and limited communication by enhancing situational awareness and modeling accuracy.
- The approach leverages a market-based coordination system, a distributed world-model estimation scheme, and a novel asymmetric entity-model (AM) integrated with Elliptical Line Voronoi Diagrams (ELVDs).
- This framework enables agents to maintain a coherent world representation and perform efficient task assignments even with high communication loss, reducing task overlaps in challenging scenarios like RoboCup soccer.
Risk-Bounded Multi-Agent Visual Navigation via Dynamic Budget Allocation
- RB-CBS (Risk-Bounded Conflict-Based Search): introduces a novel multi-agent pathfinding framework that dynamically allocates and adjusts user-specified risk bounds across agents to balance safety and efficiency in visual navigation tasks.
- The framework employs a two-level search, with a high-level component managing global constraints and risk budgets, and a low-level component finding individual agent paths within assigned risk limits.
- It integrates learned distance and risk critics from goal-conditioned reinforcement learning to construct an irregular waypoint graph, enabling collision-free path planning in visually rich, unstructured environments.
EnvX: Agentize Everything with Agentic AI
- EnvX: introduces a novel LLM-based framework that agentizes GitHub repositories into intelligent, autonomous agents capable of natural language interaction and inter-agent collaboration.
- The framework operates through three phases: TODO-guided environment initialization, human-aligned agentic automation, and Agent-to-Agent (A2A) communication, transforming static code into active, interactive components.
- EnvX integrates specialized tools and an A2A protocol to automate repository understanding, initialization, and operationalization, enabling multi-repository collaboration and enhancing software reuse.
DISENTANGLING INTERACTION AND BIAS EFFECTS IN OPINION DYNAMICS OF LARGE LANGUAGE MODELS
- Bayesian framework: introduces a Bayesian framework to disentangle and quantify three biases (topic bias, agreement bias, anchoring bias) and interaction effects in LLM opinion dynamics, where it models observed opinion shifts in multi-step dialogues between LLM agents.
- The framework quantifies the influence of these factors on opinion trajectories, revealing that LLMs tend to converge to a shared attractor, with interaction effects fading over time and biases differing between LLMs.
- It also introduces opinion uncertainty, measured by Shannon entropy, as a predictor for subsequent opinion shifts and demonstrates that fine-tuning LLMs can shift opinion attractors, highlighting both opportunities and challenges in using LLMs as proxies for human behavior.
RAFFLES: Reasoning-based Attribution of Faults for LLM Systems
- RAFFLES (Reasoning-based Attribution of Faults for LLM Systems): introduces an iterative evaluation architecture for multi-component LLM systems, utilizing a central Judge and specialized Evaluators to systematically identify decisive faults within a system's execution Trajectory (T), storing evaluation history in a Memory component (H) to refine fault attribution and output the Decisive fault (i, t).
- The framework's Judge proposes candidate agent-step fault pairs and provides rationales based on primacy, fault condition, and causality, while multiple Evaluators rigorously verify these rationales and the consistency of the proposed fault with the execution log, returning confidence scores.
- This iterative refinement process, where Evaluator feedback is fed back to the Judge via the Memory component, enables RAFFLES to achieve higher accuracy in pinpointing the earliest causal faults in complex, long-horizon LLM agentic systems compared to existing methods.
Probabilistic Modeling of Latent Agentic Substructures in Deep Neural Networks
- Probabilistic Modeling of Latent Agentic Substructures in Deep Neural Networks: introduces a theoretical framework for modeling neural agents as probabilistic generative models composed of interacting subagents, defining their beliefs and welfare through outcome distributions and epistemic utility, and aggregating them via logarithmic pooling to form a coherent composition belief.
- The framework establishes that unanimously beneficial compositions are possible with three or more outcomes under logarithmic pooling, but impossible for binary outcomes or under linear pooling, and demonstrates recursive and robustness properties for compositional agents.
- The paper formalizes the Waluigi effect in LLMs, showing that manifesting and then suppressing an antagonistic persona (Waluigi) yields greater misalignment reduction than pure reinforcement of a benevolent persona (Luigi) alone, offering insights into agentic alignment challenges.
Demo: Healthcare Agent Orchestrator (HAO) for Patient Summarization in Molecular Tumor Boards
- HAO (Healthcare Agent Orchestrator): introduces a modular, LLM-driven multi-agent system that coordinates specialized agents, a general reasoner, and domain-specific tools over unified data sources via a user interface to generate patient summaries for Molecular Tumor Boards.
- TBFact, a "model-as-a-judge" framework, evaluates patient summaries by extracting clinical factual claims, classifying their importance, assessing bidirectional entailment using an LLM, and attributing errors to quantify completeness and succinctness.
- The HAO framework is designed for precision, traceability, and safety-by-design, enabling grounded reasoning across heterogeneous data sources and supporting diverse use cases from rapid single-agent timelines to complex multi-agent workflows.
Simulating Dispute Mediation with LLM-Based Agents for Legal Research
- AgentMediation: introduces an LLM-based agent framework for simulating legal dispute mediation, featuring data preprocessing, a five-stage mediation simulation framework, configurable party and mediator agents, and a dual evaluation system.
- The framework leverages real-world civil dispute data to create structured inputs, models disputant behaviors using TKI conflict modes, and allows mediators to access external legal knowledge, enabling controlled experimentation on key variables.
- Its dual evaluation system assesses both mediation outcomes (success rate, satisfaction, consensus, litigation risk via LLM-as-a-judge) and solution quality (points of contention, legal bases using ROUGE-L, BERTScore, Recall), providing a comprehensive platform for legal research.
WebEXPLORER: Explore and Evolve for Training Long-Horizon Web Agents
- WebEXPLORER: introduces a systematic data generation approach, combining Model-Based Exploration (constructs information space) and Iterative Query Evolution (increases query difficulty), to create the WebEXPLORER-QA Dataset (synthesized challenging QA pairs) for training the WebEXPLORER-8B (trained web agent) via Supervised Fine-tuning (initializes model capabilities) and Reinforcement Learning (optimizes reasoning strategies).
- The framework generates challenging query-answer pairs that require multi-step reasoning and complex web navigation, enabling the development of advanced web agents equipped with Search Tool (retrieves relevant information) and Browse Tool (analyzes URL content).
- The WebEXPLORER-8B model, trained with this approach, achieves state-of-the-art performance at its scale on various information-seeking benchmarks and demonstrates strong generalization capabilities for long-horizon web agents.
Scaling up Multi-Turn Off-Policy RL and Multi-Agent Tree Search for LLM Step-Provers
- BFS-Prover-V2: introduces a system designed to address training-time reinforcement learning and inference-time compute scaling challenges for LLM step-provers, achieving state-of-the-art results on formal mathematics benchmarks.
- The system features a novel multi-turn off-policy RL framework with a multi-stage expert iteration pipeline, including adaptive tactic-level data filtering and periodic retraining, to continuously improve LLM performance at training time.
- It also incorporates a planner-enhanced multi-agent search architecture for inference-time scaling, where a high-level LLM Planner decomposes complex theorems into subgoals for parallel Prover agents, leveraging a shared proof cache.
- TOA (Tree of Agents): introduces a multi-agent reasoning framework that segments long input texts into chunks, processed by independent agents, to address long-context challenges like "lost in the middle" by enabling multi-perspective understanding through tree-structured path exploration.
- The framework enhances processing efficiency by incorporating prefix-hash caching to reduce redundant cognition generations and adaptive pruning strategies to terminate useless reasoning paths early.
- TOA's agents collaborate by exchanging local cognition and forming a consensus through a two-tier hierarchical voting mechanism, demonstrating comparable performance to larger commercial LLMs using a smaller base model.
Context-Adaptive Hearing Aid Fitting Advisor through Multi-turn Multimodal LLM Conversation
- CAFA (Context-Adaptive Fitting Advisor): introduces a multimodal, multi-agent LLM system for real-time, personalized hearing aid adjustments, integrating live ambient audio, audiograms, and user feedback through its Ambient Sound Recognition Pipeline, Multi-Agent Multi-turn Workflow, and LLM Judge.
- The system's Multi-Agent Multi-turn Workflow comprises a Context Acquisition Agent, Subproblem Classifier Agent, Strategy Provider Agent, and Ethical Regulator Agent, all overseen by an independent LLM Judge to ensure clinical safety and quality.
- CAFA leverages LLMs for multi-step reasoning and agentic task execution, translating context and user feedback into precise, safe tuning commands, and achieving high ambient sound classification accuracy for enhanced conversational efficiency.
Evaluating Multi-Turn Bargain Skills in LLM-Based Seller Agents
- BargainBench framework: introduces a multi-turn evaluation framework for LLM-based seller agents, including an Intent Factory (extracts intent space), Problem Weaver (generates scripted dialogues), and Evaluation Center (scores LLM performance), designed to measure bargaining ability by tracking buyer intents in e-commerce dialogues.
- The framework provides a large-scale e-commerce bargaining benchmark with turn-level evaluation grounded in Theory of Mind, moving beyond outcome-only metrics to assess intermediate reasoning.
- It also features an automated pipeline for extracting reliable intent from massive dialogue data, enabling scalable and reproducible benchmarking of bargaining agents.
A Fragile Number Sense: Probing the Elemental Limits of Numerical Reasoning in LLMs
- Divide-and-Reconstruct Framework: introduces a multi-level framework that decomposes complex numerical reasoning tasks into elementary skills, evaluating LLMs on both isolated skills and their integration to analyze performance and identify reasoning limitations.
- The paper probes LLM mathematical numeracy across escalating complexity, from basic arithmetic to combinatorial puzzles like the Game of 24, revealing a "fragile number sense" in LLMs.
- Results indicate LLMs excel at deterministic algorithmic execution but consistently fail at tasks requiring heuristic search over large combinatorial spaces, suggesting their numerical reasoning is more pattern-matching than generative problem-solving.
SFR-DeepResearch: Towards Effective Reinforcement Learning for Autonomously Reasoning Single Agents
- SFR-DeepResearch: introduces a framework for training autonomous single-agent LLMs for Deep Research, featuring an Agentic Inference Pipeline with a Memory Management System and essential tools, and an RL Training Recipe that includes a Synthetic Data Generation Pipeline, a modified REINFORCE Algorithm, and a Reward Modeling System.
- The framework enhances agentic capabilities by continually training reasoning-optimized LLMs using synthetic data and a novel RL recipe that incorporates length-normalized advantage and strategic trajectory filtering to stabilize policy optimization.
- The system also includes a robust RL Infrastructure with asynchronous processing, a local toolbox, and optimized GPU resource management to ensure scalability and fault-tolerance during training.
TableMind: An Autonomous Programmatic Agent for Tool-Augmented Table Reasoning
- TableMind (Autonomous Programmatic Agent for Tool-Augmented Table Reasoning): introduces an LLM-driven table reasoning agent that autonomously performs multi-turn tool invocation, writes and executes data-analyzing code in a secure sandbox, and exhibits planning and self-reflection capabilities, utilizing a two-stage fine-tuning paradigm with SFT and RFT, enhanced by RAPO.
- The agent operates through an iterative plan-action-reflect loop, where the Planning Component formulates strategies, the Action Component generates Python code executed in a Code Sandbox/Interpreter, and the Reflection Component evaluates outcomes to dynamically adjust subsequent actions.
- Training involves a Supervised Fine-tuning (SFT) Module for foundational tool-use patterns and a Reinforcement Fine-tuning (RFT) Module, guided by a Multi-objective Reward Function and the Rank-Aware Policy Optimization (RAPO) Algorithm, to achieve strategic autonomy and robust problem-solving.
REMI: A Novel Causal Schema Memory Architecture for Personalized Lifestyle Recommendation Agents
- REMI (Causal Schema Memory): introduces a novel architecture for personalized multimodal lifestyle agents that combines a personal causal knowledge graph, a causal reasoning engine, a schema-based planner, and LLM orchestration to deliver explainable, personalized recommendations.
- The framework leverages two LLM components: an LLM (Reasoning) for path scoring and hypothesis generation, and an LLM (Orchestrator) for integrating information and composing final recommendations.
- This approach addresses limitations of current personalized recommendation agents by providing context-aware, user-aligned, and transparent causal explanations for lifestyle advice.
Neuro-Symbolic AI for Cybersecurity: State of the Art, Challenges, and Opportunities
- G-I-A (Grounding-Instructibility-Alignment) Framework: introduces, "a novel framework to evaluate Neuro-Symbolic AI (NeSy) systems in cybersecurity", with Grounding (connects outputs to cybersecurity concepts), Mathematical Consistency (aligns neural predictions with logical reasoning), Knowledge Graphs (provides domain knowledge), Adversarial Robustness (maintains stability against novel attacks), Instructibility (responds to analyst feedback), Dynamic Adaptation (adapts to evolving threats), Human-AI Collaboration (integrates expert guidance), Analyst Feedback (updates system components), Alignment (ensures consistency with cybersecurity objectives), Institutional Goals (aligns with organizational priorities), Ethical Constraints (prevents malicious misuse), Cybersecurity Objectives (reflects true security goals), Neural Component (performs pattern recognition), Symbolic Component (provides domain knowledge and rules), Neuro-Symbolic Fusion (integrates neural and symbolic outputs), Explainable Alerts (generates human-readable security alerts), where this framework systematically characterizes the field by analyzing 127 publications spanning 2019-July 2025.
- The paper highlights consistent advantages of multi-agent NeSy architectures, identifies critical implementation challenges, and emphasizes causal reasoning integration for proactive defense strategies.
- The survey demonstrates dual-use implications where autonomous systems achieve substantial zero-day exploitation capabilities with significant cost reductions, fundamentally altering threat dynamics and demanding responsible development.
AxelSMOTE: An Agent-Based Oversampling Algorithm for Imbalanced Classification
- AxelSMOTE (Axelord Synthetic Minority Oversampling Technique): introduces an agent-based oversampling algorithm for imbalanced classification, treating data instances as autonomous agents, and includes trait-based feature grouping (partitions features into related groups), similarity assessment (calculates cultural similarity between agents), a probabilistic exchange mechanism (exchanges traits based on similarity and probability), Beta blending (interpolates features realistically), and diversity injection (adds controlled noise for variety).
- This approach addresses limitations of traditional oversampling by preserving feature correlations, ensuring meaningful interactions between compatible instances, and generating diverse synthetic samples to avoid overfitting.
- Inspired by Axelrod's cultural dissemination model, AxelSMOTE systematically generates realistic synthetic minority instances, outperforming state-of-the-art methods while maintaining computational efficiency.
Agentic DDQN-Based Scheduling for Licensed and Unlicensed Band Allocation in Sidelink Networks
- Agentic DDQN-based scheduling framework: introduces an AI-driven DDQN framework for licensed and unlicensed band allocation in NR sidelink networks, comprising a DDQN Agent, Online Network, Target Network, Replay Buffer, Scheduler, State Representation, Action Space, and Reward Mechanism, designed to autonomously perceive network dynamics and adapt scheduling policies.
- This framework leverages agentic AI principles to integrate queueing delay, channel quality, Wi-Fi coexistence dynamics, and link-switching stability into its state representation and reward function, enabling QoS-aware and adaptive resource allocation.
- The proposed system significantly reduces blocking rates by up to 87.5% compared to threshold-based scheduling, demonstrating its potential for stable and adaptive resource management in congested and coexistence-limited environments.
MAS-Bench: A Unified Benchmark for Shortcut-Augmented Hybrid Mobile GUI Agents
- MAS-Bench (Unified Benchmark for Shortcut-Augmented Hybrid Mobile GUI Agents): introduces a benchmark for evaluating GUI-shortcut hybrid mobile agents, featuring a Shortcut Generation Stage (agent creates shortcuts), a Quality Evaluation Stage (evaluates generated shortcuts), an Online Evaluation Environment (dynamic Android platform), a GUI-Shortcut Hybrid Action Space (combines GUI and shortcut actions), Tasks Design (complex real-world scenarios), and Evaluation Metrics (Success Rate, Efficiency, Cost), where it systematically assesses agents' ability to discover, generate, and utilize shortcuts to enhance task efficiency and success.
- The benchmark includes 139 complex tasks across 11 real-world applications, a knowledge base of 88 predefined shortcuts (APIs, deep-links, RPA scripts), and supports the dynamic generation of new shortcuts by agents.
- Experiments demonstrate that hybrid agents achieve significantly higher success rates and efficiency than GUI-only counterparts, highlighting the effectiveness of integrating shortcuts and the potential for agent-generated shortcuts.
Interactive Shaping of Granular Media Using Reinforcement Learning
- RL framework: introduces an RL framework that enables a robotic arm with a cubic end-effector and a stereo camera to shape granular media into desired target structures, utilizing compact observations and concise reward formulations for effective learning.
- This framework reconstructs the current height map (Current He) from depth images (Depth image Ic) and uses its difference (Heightmap difference Ha) with the desired goal height map (Goal Hg) as a key observation for the Policy π (RL Agent).
- The approach demonstrates zero-shot transfer of trained policies from physics simulation to a real robot, outperforming baselines in shaping accuracy and robustness for complex granular media manipulation tasks.
STAYING IN THE SWEET SPOT: RESPONSIVE REASONING EVOLUTION VIA CAPABILITY-ADAPTIVE HINT SCAFFOLDING
- SEELE (reSponsive rEasoning Evolution via capability-adaptivE hint scaffolding): introduces a novel supervision-aided Reinforcement Learning with Verifiable Rewards (RLVR) framework that dynamically adjusts problem difficulty to maintain high learning efficiency by adapting hint lengths based on the LLM's evolving capability.
- This framework employs a multi-round rollout sampling strategy and an Item Response Theory-based accuracy-hint model to predict and achieve an optimal 50% rollout accuracy for each problem instance.
- By integrating real-time feedback and instance-level difficulty adjustment, SEELE significantly outperforms existing RLVR methods on various math reasoning and general domain benchmarks.
DCPO: Dynamic Clipping Policy Optimization
- DCPO (Dynamic Clipping Policy Optimization): introduces a novel reinforcement learning pipeline that enhances LLM reasoning capabilities through its Dynamic Adaptive Clipping (DAC) mechanism, which adaptively adjusts clipping bounds based on token-specific prior probabilities, and its Smooth Advantage Standardization (SAS) technique, which standardizes advantages across cumulative training steps, further incorporating an Only Token Mean loss (OTM) for efficient gradient updates.
- The framework addresses limitations of existing policy optimization methods, such as zero gradients and inefficient data utilization, by promoting token-level exploration and stabilizing training.
- DCPO demonstrates superior performance on mathematical reasoning benchmarks, significantly improving response utilization, training efficiency, and reducing token clipping compared to baselines like GRPO and DAPO.
Instruction Agent: Enhancing Agent with Expert Demonstration
- Instruction Agent: introduces a GUI agent leveraging expert demonstrations to generate step-by-step instructions, which are then executed by an Actor model incorporating verification and backtracking for robustness.
- The framework, composed of an Instructor (Recorder, Instruction Generator) and an Actor (Grounder, Executor, Verifier, Backtracker), automates complex GUI tasks by strictly following demonstrated trajectories.
- Utilizing LLMs (GPT-40) for planning, verification, and execution, and UI-Tars 1.5 for grounding, the agent achieves a 60% success rate on tasks previously failed by top-ranked GUI agents.
- LLM-DCMF (Large-Language-Model-Driven Composite Materials Research Framework): introduces a language-native, lightly structured database and LLM-based data management system, with Light-Structured Text Generation (processes raw scientific literature into modular, lightly structured text using LLM-guided prompts), Heterogeneous Database (stores and organizes both lightly structured text units and fully structured data derived from them), Hybrid Search (retrieves relevant information from the database using multiple methods), and Application Layer (RAG & Agentic Workflows) (utilizes retrieved information for reasoning, generation, and iterative design), where the system transforms raw scientific literature into a queryable, heterogeneous database to accelerate materials discovery.
- The framework captures lightly structured information from papers across preparation, characterization, theory/computation, and mechanistic reasoning, organizing records in a heterogeneous database for composite retrieval with semantics, keywords, and value filters.
- This system synthesizes literature into accurate, verifiable, and expert-style guidance, enabling high-fidelity Retrieval-Augmented Generation (RAG) and tool-augmented agents to interleave retrieval with reasoning for actionable SOPs.
Proof2Silicon: Prompt Repair for Verified Code and Hardware Generation via Reinforcement Learning
- Proof2Silicon: introduces an end-to-end synthesis framework that leverages PREFACE's RL-driven prompt optimization core to guide a frozen LLM in generating formally verified Dafny code, which is then translated through a PyLog-based pipeline and Vivado HLS into synthesizable RTL for FPGA hardware.
- The framework integrates a verifier-guided RL agent (SLM) to iteratively refine prompts based on formal verification feedback, ensuring Dafny code correctness without costly LLM fine-tuning.
- This pipeline bridges natural language specifications with silicon realization, enabling automated, correctness-by-construction hardware synthesis for safety-critical domains.
PillagerBench: Benchmarking LLM-Based Agents in Competitive Minecraft Team Environments
- TactiCrafter: introduces an LLM-based multi-agent system for competitive Minecraft environments, featuring a Tactics Module (generates high-level strategies), a Causal Model (learns causal relationships), an Opponent Model (infers enemy strategies), and Base Agents (execute actions, self-reflect).
- This system facilitates teamwork through human-readable tactics, learns causal dependencies from gameplay, and adapts to opponent strategies through repeated self-play.
- TactiCrafter is evaluated on PillagerBench, a novel benchmark for real-time competitive team-vs-team Minecraft scenarios, demonstrating superior performance over baselines and adaptive learning.
- LLM Diary Generation Workflow: introduces a novel agent-based framework for generating individual travel diaries using LLMs, leveraging components like a Block Group Data Loader, Persona Synthesis Stage, Prompt Engineering Stage, LLM Call (llama3), and Generation & Parsing Stage to produce realistic travel patterns.
- This two-stage framework first stochastically synthesizes demographically consistent personas from open-source census and land-use data, then uses a comprehensive prompt to direct the LLM (Llama 3) to generate a full day's structured travel diary.
- The framework employs a rigorous one-to-cohort validation strategy, comparing LLM-generated diaries against real-world survey data and classical models, demonstrating the LLM's zero-shot viability and superior semantic understanding for travel behavior.
- Entropy-Driven Dialogue Policy (EDDP): introduces a novel method for conversational recommender systems to balance exploration and exploitation by dynamically routing dialogue based on the entropy of retrieval score distributions.
- This policy quantifies user intent broadness via a broadness score derived from re-ranked product retrieval scores, enabling the system to ask clarifying questions for high-entropy (ambiguous) queries or make direct recommendations for low-entropy (precise) queries.
- Integrated within a multi-skill e-commerce AI agent, the EDDP allows an LLM-driven system to adapt to arbitrarily large product catalogs in real-time without context window bloat, enhancing shopper engagement.
- Experimental Design for Credence Goods Markets: introduces a series of one-shot experiments to quantify the behavioral, welfare, and distribution consequences of LLMs in expert service markets, varying market interaction types, institutional environments, LLM objective functions, training regimes, and transparency rules.
- The study finds that Human-Human markets generally achieve higher efficiency than AI-AI and Human-AI markets due to pro-social expert preferences and higher consumer trust, while LLM experts often earn higher surplus at the expense of consumers.
- Crucially, allowing human experts to delegate to LLMs and codify their LLM's social preferences, especially when transparent to consumers, significantly increases market efficiency and reduces fraud, potentially outperforming Human-Human markets under transparency rules.
- PolicyEvolve: introduces a general framework for generating programmatic policies in multi-player games, with Global Pool (preserves elite policies), Local Pool (stores temporary policies), Policy Planner (generates/refines policy code), and Trajectory Critic (analyzes policy performance).
- The framework leverages LLMs for policy generation and iterative refinement, using population-based training to evolve robust and interpretable rule-based policies.
- PolicyEvolve significantly reduces reliance on manually crafted policy code, achieving high-performance policies with minimal environmental interactions and demonstrating consistent strategy evolution.
- Code2MCP: introduces an automated, multi-agent framework for transforming GitHub repositories into Model Context Protocol (MCP) services, employing LLM-powered agents for code analysis, generation, review, and finalization.
- The framework operates via a state graph orchestrating a multi-stage workflow, featuring a closed-loop "Run-Review-Fix" cycle for autonomous debugging and repair of generated code.
- Code2MCP significantly accelerates the MCP ecosystem by systematically converting open-source code into deployable services and comprehensive technical documentation with minimal human intervention.
MapAgent: A Hierarchical Agent for Geospatial Reasoning with Dynamic Map Tool Integration
- MapAgent: introduces a hierarchical multi-agent plug-and-play framework for map-integrated geospatial reasoning, featuring a Planner Agent, Module Inventory (Visual Place Recognizer, Map Service, Sequencer, Solution Generator, Answer Generator), and a Map-Tool Agent with specialized map tools (Nearby, Trip, Route, PlaceInfo).
- This framework decouples high-level planning from low-level tool execution, allowing the Planner Agent to decompose complex queries into subgoals routed to specialized modules, while the Map-Tool Agent adaptively orchestrates map APIs.
- MapAgent significantly outperforms existing tool-augmented and agentic baselines across diverse geospatial benchmarks, demonstrating robust performance and generalizability in real-world map-integrated reasoning tasks.
Let's Roleplay: Examining LLM Alignment in Collaborative Dialogues
- FAAF (Frictional Agent Alignment Framework): introduces a novel counterfactual evaluation framework to examine LLM alignment in collaborative dialogues, utilizing a roleplay methodology with a Friction Agent (πF) and Collaborator Agent (πC) to study the impact of friction interventions on common ground and task outcomes.
- The framework employs an Oracle Agent (O) for data generation, simulating frictive states and interventions, and evaluates performance using metrics like Common Ground (CG) size, solution accuracy, and a Reward Model (RM) for intervention quality.
- This approach addresses the suboptimality risks in multi-party interactions by explicitly modeling action modifications within a Modified-Action MDP (MAMDP) and demonstrates that friction-aware alignment improves both common ground convergence and task correctness.
Chatbot To Help Patients Understand Their Health
- NoteAid-Chatbot (Learning as Conversation Framework): introduces a conversational AI system for patient education, built on a multi-agent LLM and reinforcement learning setup, which leverages synthetic data for supervised fine-tuning and PPO-based alignment with rewards derived from simulated patient comprehension assessments.
- This framework automates training without human-labeled data, enabling the development of a lightweight, domain-specific chatbot capable of multi-turn interactions and diverse educational strategies.
- Evaluations, including a Turing test, demonstrate the system's ability to surpass non-expert human educators in patient understanding and exhibit key emergent behaviors like clarity and structured dialogue.
DRF: LLM-AGENT Dynamic Reputation Filtering Framework
- DRF (LLM-AGENT Dynamic Reputation Filtering Framework): introduces a multi-agent system that constructs an interactive rating network, designs a reputation scoring mechanism, and integrates an Upper Confidence Bound-based strategy to quantify agent performance and enhance selection efficiency.
- The framework utilizes a core agent for decision-making and control, while task agents execute and evaluate subtasks, with their performance and credibility dynamically assessed through a k-layer rating network and iterative reputation updates.
- DRF significantly improves task completion quality and collaboration efficiency in complex tasks like logical reasoning and code generation by prioritizing high-reputation, low-cost LLM agents.
Exploit Tool Invocation Prompt for Tool Behavior Hijacking in LLM-Based Agentic System
- TEW (TIP Exploitation Workflow): introduces a systematic approach to identify and exploit vulnerabilities in LLM-based agentic systems, comprising Prompt Stealing (extracts prompt components), TIP Vulnerabilities Analysis (reconstructs TIP and identifies weaknesses), and TIP Hijacking (exploits tool invocation).
- This framework demonstrates how manipulating Tool Invocation Prompts (TIPs) can lead to severe security risks such as remote code execution (RCE) and denial of service (DoS) in LLM-based agentic systems.
- The research highlights the critical need for robust security measures in TIPs, as current defense mechanisms, including guard models and self-reflection, are often insufficient against sophisticated attacks.
A Composable Agentic System for Automated Visual Data Reporting
- Composable Agentic System: introduces a multi-stage, hybrid agentic architecture for automated visual data reporting, leveraging LLM-driven agents for reasoning and programmatic/rule-based components for deterministic logic, producing interactive Observable 2.0 reports and executable Marimo notebooks.
- The system's workflow is coordinated by an Orchestrator and includes distinct phases for data understanding, analysis and materialization, visualization, and reporting, with components like Field Refiner, Dataset Profiler, Insight Planner, Dataset Visualizer, and Report Narrator.
- This architecture supports a Human-AI Partnership Model by providing auditable, steerable, and granular outputs, externalizing critical logic to deterministic modules like the Draco visualization system, and enabling deep traceability for analysts and interactive exploration for readers.
Orchestrator: Active Inference for Multi-Agent Systems in Long-Horizon Tasks
- Orchestrator: introduces a novel multi-agent system (MAS) framework that leverages attention-inspired self-emergent coordination and reflective benchmarking to optimize global task performance, featuring a modular cell architecture with planning, execution, and orchestration nodes.
- The framework integrates a monitoring mechanism to track agent-environment dynamics and agent-to-agent interaction, using active inference benchmarks to optimize system behavior and mitigate partial observability.
- Orchestrator dynamically adjusts LLM agents' internal policies and prompt designs based on performance metrics, guiding them away from local solution minima towards global task objectives in complex, long-horizon environments.
ProfilingAgent: Profiling-Guided Agentic Reasoning for Adaptive Model Optimization
- ProfilingAgent: introduces a profiling-guided agentic approach that automates model compression through structured pruning and post-training dynamic quantization, leveraging a multi-agent system including an Acquisition Agent, Input Shape Resolver Agent, Profiling Agent, Analysis Agent, Pruning Agent, Quantization Agent, Evaluation Agent, and Iterative Pruning Agent.
- This modular pipeline utilizes LLM-guided agents, specifically the Input Shape Resolver Agent and Analysis Agent, to interpret static and dynamic profiling signals, generating architecture-specific compression strategies and layer-wise decisions tailored to performance bottlenecks.
- The system adaptively refines pruning decisions through iterative feedback, balancing accuracy and latency, and demonstrates superior performance compared to heuristic baselines across various vision models.
FAIR RISK OPTIMIZATION OF DISTRIBUTED SYSTEMS
- FRODS (Fair Risk Optimization of Distributed Systems): introduces a framework for assessing and optimizing the total risk of complex distributed systems, utilizing systemic risk measures and a decomposition method to solve a two-stage stochastic programming problem for fair risk allocation.
- The framework addresses challenges like non-additive risk, confidential information, and fair risk allocation by evaluating individual agent risks and aggregating them into a total system risk measure.
- The decomposition method enables distributed decision-making, allowing agents to operate autonomously with minimal information exchange while contributing to overall system risk minimization.
ArGen: Auto-Regulation of Generative AI via GRPO and Policy-as-Code
- ArGen (Auto-Regulation of Generative AI systems): introduces a novel framework for aligning LLMs with complex, configurable, machine-readable rules by synthesizing principle-based automated reward scoring, Group Relative Policy Optimisation (GRPO), and an OPA-inspired governance layer.
- This framework operationalizes nuanced ethical principles and regulatory compliance standards into programmable reward functions and explicit policies, enabling continuous auto-regulation and verifiable compliance.
- ArGen's architecture allows for live policy hot-swapping without model retraining, providing a transparent, auditable, and adaptable path to governable AI systems in diverse global contexts.
- LatticeWorld: introduces a multimodal Large Language Model-Empowered Framework for Interactive Complex World Generation, with User Inputs (textual/visual instructions), LLML (Large Language Model for Layout Generation) (generates symbolic layout), Visual Encoder (Φ) (extracts visual features), Projection Module (Proj) (maps visual features), LLMc (Large Language Model for Environmental Configuration) (generates environment configurations), Decoder (ΨL) (translates symbolic layout), Configuration Translator (ΨC) (interprets environment configurations), Rendering Engine (Unreal Engine 5) (creates dynamic 3D world), and PCG Pipeline (Procedural Content Generation Pipeline) (automates asset creation), where it generates large-scale 3D interactive worlds with dynamic agents and high-fidelity physics simulation from multimodal user instructions.
- The framework leverages lightweight LLMs (LLaMA-2-7B) and an industry-grade rendering engine (Unreal Engine 5) to process textual descriptions and visual instructions (e.g., height maps) into a symbolic layout representation and environmental configurations.
- LatticeWorld achieves over a 90x increase in industrial production efficiency for 3D environment generation compared to traditional manual methods, while maintaining high creative quality and enabling real-time multi-agent interaction and simulation.
- TAXAL (Triadic Alignment for eXplainability in Agentic LLMs): introduces a triadic fusion framework for explainable LLMs, integrating Cognitive (user understanding, intelligibility), Functional (practical utility, compliance), and Causal (faithful reasoning, intervention) dimensions to provide a unified, role-sensitive foundation for designing, evaluating, and deploying explanations in diverse sociotechnical settings.
- The framework aims to enhance Trust (user confidence), Contestability (challenge outputs), and Accountability (auditability, compliance) in agentic LLMs by aligning explanation strategies with human cognition, institutional expectations, and enabling users to challenge outputs and audit traces.
- TAXAL provides a step-by-step guide for identifying Stakeholder Role Identification (define audience), selecting Relevant Dimension Selection (choose method), choosing Explanation Strategy Selection (choose method), balancing Trade-Off Balancing (optimize dimensions), and iteratively refining explanations through Contextual Iteration (evaluate, refine), as demonstrated through cross-domain case studies.
AI Agents for Web Testing: A Case Study in the Wild
- WebProber: introduces an AI agent-based web testing framework that autonomously explores websites, simulates user interactions, identifies bugs, and generates human-readable reports, comprising a prompt generation module (generates testing prompts), an interaction module (simulates user experience), and a report generation module (analyzes interaction history), leveraging a Visual Language Model (VLM) and a bug database.
- The framework interacts directly with visual webpages using VLMs to simulate human-like behavior, enabling the detection of contextual usability issues and bugs often overlooked by traditional web testing tools.
- The system's case study on 120 academic personal websites demonstrated its ability to uncover subtle, human-centric problems, highlighting the potential of agent-based testing while also revealing challenges in agent-browser interaction and bug coverage.
Shared Autonomy through LLMs and Reinforcement Learning for Applications to Ship Hull Inspections
- Shared Autonomy Framework: introduces a multi-layered architecture for ship hull inspections, integrating a Supervisor (LLM), a Mission Manager (Behavior Trees), and a Multi-Agent Execution Layer (DRL-trained Agents) to enable human-robot collaboration.
- The framework allows an Operator to specify high-level goals via natural language, which are translated into structured tasks managed by Behavior Trees, while DRL-trained agents execute tasks and adapt their behavior in a leader-follower configuration.
- This system aims to reduce operator cognitive load, enhance transparency, and improve adaptive behavior alignment with human intent in complex maritime environments.
- Dual-Loop MAS (Dual-Loop Multi-Agent System with Parallel Terminal-Edge Collaboration): introduces an LLM-enabled multi-agent system for 6G networks, featuring a Global Agent, Sub-Agent, Perception Module, Planning Module (Outer Loop / Inner Loop), LLMCompiler, Memory Module, Scheduling Module, Tool Execution, Database, Self-Evolution, and Knowledge Management, designed to enhance task planning and execution efficiency through dual-loop edge-terminal collaboration.
- The framework utilizes an outer loop for global task decomposition by a Global Agent and an inner loop where Sub-Agents, guided by LLMCompiler, perform parallel sub-task execution and replanning with tool calling and offloading strategies.
- It integrates a comprehensive Memory Module for context and knowledge, a Perception Module for multimodal information and intention recognition, and a Scheduling Module for efficient resource allocation, enabling self-evolution and robust knowledge management in 6G environments.
Internet 3.0: Architecture for a Web-of-Agents with it's Algorithm for Ranking Agents
- DOVIS (Discovery, Orchestration, Verification, Incentives, Semantics) and AgentRank-UC: introduces a framework for the Agentic Web, with DOVIS as a five-layer operational protocol for collecting privacy-preserving telemetry and AgentRank-UC as a dynamic, trust-aware algorithm for ranking agents based on usage and competence.
- The DOVIS protocol includes Discovery, Orchestration, Verification, Incentives, and Semantics layers, enabling verifiable and incentivized telemetry collection through the OAT-Lite schema.
- AgentRank-UC combines usage and competence signals into a unified ranking, ensuring scalability, trustworthiness, and resilience against manipulation in open agent ecosystems.
Towards Ontology-Based Descriptions of Conversations with Qualitatively-Defined Concepts
- Ontology-Based Conversational Control Framework: introduces an approach for controlling conversational generation by quantitatively defining qualitatively-described conversation aspects, with Qualitatively-defined conversational features (high-level concepts), Pre-defined descriptors (measurable text properties), Annotated dataset (labeled text examples), Classifier (maps descriptors to concepts), Subclass membership rules (defines concept boundaries), Description logic (formal rule representation), Ontology (structured knowledge base), Reasoning step (infers new knowledge), LLM fine-tuning step (adapts LLM behavior), LLM (generates controlled text), and LoRA adapters (efficiently tunes LLM).
- The framework leverages linguistic descriptors and a Decision Tree Classifier to establish structured definitions of conversational features, which are then integrated into an ontological framework for consistency and transparency.
- This methodology enables the fine-tuning of LLMs, such as Llama3-8B-Instruct, to generate content that adheres to specific ontological concepts, demonstrated through a Proficiency-Level Control use-case.
OSC: Cognitive Orchestration through Dynamic Knowledge Alignment in Multi-Agent LLM Collaboration
- OSC (Orchestrating Cognitive Synergy): introduces a knowledge-aware adaptive collaboration framework designed to enhance cognitive synergy in multi-agent LLM systems, with its Expert Pool (collection of LLM agents), Query (initial task/question), Collaborator Knowledge Model (dynamically tracks collaborators' cognitive states), Cognitive Gap Analysis Module (identifies discrepancies between agents), Communication Policy (shapes communication behavior), Linguistic Realization Engine (translates abstract action to language), Dialogue History (record of past interactions), Internal State (agent's own cognitive state), Update Mechanism (updates CKM based on dialogue), Aggregator Module (combines individual responses), Reward Function (guides learning), Task Performance Reward (feedback for task success), Communication Cost (penalty for message length), Intrinsic Shaped Reward (guides collaborative behaviors), and Reinforcement Learning (optimizes communication policy).
- The framework operates as an intermediate layer between expert selection and aggregation, enabling agents to dynamically perceive collaborators' cognitive states, analyze cognitive gaps, and adapt communication behaviors using learned strategies.
- This dynamic approach transforms parallel-working LLM agents into a deeply collaborative cognitive team, significantly improving task performance and communication efficiency in complex reasoning and problem-solving benchmarks.
- VoltanaLLM: introduces an SLO-aware, energy-efficient LLM serving system built on prefill/decode (P/D) disaggregation, with EcoFreq (dynamically adjusts GPU frequency for prefill/decode phases), EcoRoute (routes decode requests to optimize energy), EcoPred (predicts TTFT/ITL latency), Prefill Instances (process initial LLM input), Decoding Instances (generate subsequent LLM tokens), New Requests (incoming user queries), Generated Tokens (output streamed to user), Load Metrics (real-time system performance data), TTFT prediction (Time-To-First-Token estimation), and ITL prediction (Inter-Token Latency estimation).
- This framework co-designs frequency scaling and request routing, leveraging decoupled execution to enable fine-grained, phase-specific control for LLM inference.
- VoltanaLLM achieves up to 36.3% energy savings while maintaining near-perfect SLO attainment rates across various LLMs and real-world datasets.
- One-SHOT and Iterative Retrieval Strategies: introduces two RAG approaches, One-SHOT for token-constrained retrieval with chunk filtering and cropping, and an iterative agentic framework with an LRM for multi-turn query refinement and context management.
- The paper addresses common RAG bottlenecks, such as missing crucial "golden chunks" in top-k retrieval and issues like query drift and retrieval laziness in complex QA tasks.
- These strategies offer practical solutions for improving evidence coverage and answer quality in legal and regulatory domains, particularly with heterogeneous government documents.
TalkToAgent: A Human-centric Explanation of Reinforcement Learning Agents with Large Language Models
- TalkToAgent: introduces a multi-agent LLM framework for human-centric explanation of RL agents, featuring Coordinator, Explainer, Coder, Evaluator, and Debugger agents, along with predefined XRL tools and generated codes.
- The framework interprets natural language queries, maps them to relevant XRL tools, and provides multimodal explanations including figures and domain-aware textual interpretations.
- TalkToAgent enhances counterfactual explanations by generating alternative scenarios from qualitative behavioral descriptions or new rule-based policies, ensuring high accuracy and minimizing failures through agent interactions.
Mind the Gap: Evaluating Model- and Agentic-Level Vulnerabilities in LLMs with Action Graphs
- AgentSeer: introduces an observability-based evaluation framework that decomposes agentic executions into granular actions and components, enabling systematic agentic-situational assessment of LLMs.
- The framework addresses critical gaps in current LLM safety evaluation by transforming opaque agentic executions into structured, analyzable representations through a knowledge graph.
- It empirically validates that agentic deployments have distinct "agentic-only" vulnerabilities, often invisible to traditional model-level testing, emphasizing the need for agentic-situation evaluation.
Personality as a Probe for LLM Evaluation: Method Trade-offs and Downstream Effects
- Unified Evaluation Framework: introduces a systematic study for LLM personality control, comparing In-Context Learning, Parameter-Efficient Fine-Tuning, and Mechanistic Steering methods, utilizing a Contrastive Dataset Generation, Trait Purification Techniques, and a Three-Level Stability Framework for comprehensive evaluation on Gemma-2 and LLaMA-3 models.
- The framework assesses method trade-offs in personality alignment, task performance, and demographic bias across MMLU, GAIA, and BBQ benchmarks.
- It provides practical guidance for selecting personality manipulation methods based on alignment strength, computational requirements, and performance preservation.
SePA: A Search-enhanced Predictive Agent for Personalized Health Coaching
- SePA (Search-enhanced Predictive AI Agent): introduces a novel LLM health coaching system that integrates personalized machine learning and retrieval-augmented generation to deliver adaptive, evidence-based guidance, featuring a User Interface (web application for interaction), Data Ingestion Module (uploads and processes raw data), Data Preprocessing & Feature Engineering Module (transforms raw data), Predictive Modeling Module (forecasts health risks), Toolbox Module (provides analytical capabilities), Conversational LLM Agent (manages user dialogue), and a Web Retrieval Pipeline (grounds LLM advice).
- SePA forecasts daily stress, soreness, and injury risk from wearable sensor data using a two-tiered predictive modeling strategy, employing a generalized XGBoost model for new users and personalized deep neural networks for engaged users.
- The system's web-retrieval pipeline dynamically rewrites search queries based on ML predictions and grounds advice in a curated whitelist of trusted sources, ensuring contextual relevance, verifiability, and transparency.
Cloning a Conversational Voice AI Agent from Call Recording Datasets for Telesales
- Conversational Voice AI Agent Cloning System: introduces a general methodology for cloning a conversational voice AI agent from call recordings, leveraging a cloning system to extract behavioral patterns and an inference system to deploy the agent in live calls.
- The cloning pipeline transforms raw call recordings into a structured Agent Playbook, which defines the agent's persona, knowledge, and conversational style through prompt engineering.
- The inference system deploys the cloned agent in real-time using a Voice LLM Core, which processes user speech and generates synthetic agent responses for fluid, end-to-end spoken interaction.
LEARNING TOOL-AWARE ADAPTIVE COMPLIANT CONTROL FOR AUTONOMOUS REGOLITH EXCAVATION
- LTACC (Learning Tool-Aware Adaptive Compliant Control): introduces a learning-based framework where a Model-based RL Agent learns adaptive compliance within a Space Robotics Bench (SRB) simulation, leveraging SimForge Engine and Procedural Generation to create diverse scenarios with High-fidelity Particle Physics (XPBD) for autonomous regolith excavation.
- The framework utilizes an Operational Space Control (OSC) Controller to dynamically modulate robot stiffness and damping, enabling the RL Policy to adapt to various Excavation Tools and Regolith properties.
- The system integrates Vision and Proprioception feedback, processed by an Encoder and a Recurrent World Model, to infer hidden states and achieve generalizable, tool-aware excavation skills across Parallelized Simulation environments.
Psychologically Enhanced AI Agents
- MiT (MBTI-in-Thoughts): introduces a framework for enhancing LLM agents through psychologically grounded personality conditioning, with LLM Agents (core processing units), Psychological Profiles (personality conditioning via prompts), 16Personalities Test (external validation tool), Majority Voting Protocol (isolated reasoning mechanism), Interactive Communication Protocol (decentralized dialogue mechanism), Interactive Communication With Self-Reflection Protocol (dialogue with private memory), Self-Reflection (private deliberation buffer), Blackboard (shared communication memory), and LLM as Judge (final decision maker).
- The framework primes LLM agents with distinct MBTI personality archetypes via prompt engineering, enabling control over behavior along cognitive and affective axes, and supports structured multi-agent communication protocols.
- MiT demonstrates that personality priming induces consistent behavioral biases, improves cooperation and reasoning quality through self-reflection, and generalizes to other psychological models without fine-tuning.
EvoEmo: Towards Evolved Emotional Policies for LLM Agents in Multi-Turn Negotiation
- EvoEmo framework: introduces an evolutionary reinforcement learning framework for optimizing dynamic emotional expression in multi-turn negotiations, with Negotiation Setup, Seller LLM Agent, Buyer LLM Agent, Product Description, Seller Prompts, Buyer Prompts, EvoEmo Optimization Module, Emotion States, Emotional Policy (πω), Policy Population, Population Generation, Selection Operator, Crossover Operator, Mutation Operator, Negotiation Simulation Module, Simulated Seller Agents, Simulated Buyer Agents, Mediator Agent, Evaluation Module, Reward Function (R(S)), Optimal Policy (πω*), and Termination Conditions, where it evolves high-reward emotion policies using population-based genetic optimization across diverse negotiation scenarios.
- The framework models emotional state transitions as a Markov Decision Process and iteratively refines policies based on rewards achieved during negotiations, combining evolutionary exploration with reinforcement learning principles.
- EvoEmo consistently outperforms vanilla and fixed-emotion baselines, achieving higher success rates, efficiency, and buyer savings by enabling LLM agents to adaptively express emotions.
Are LLM Agents the New RPA? A Comparative Study with RPA Across Enterprise Workflows
- AACU (Agentic Automation with Computer Use): introduces a comparative study of LLM agents and traditional Robotic Process Automation (RPA) across enterprise workflows, including LLM Agent (intelligent core), Computer Use Capability (interacts digital systems), User Interface (UI) Interaction (mimics human actions), Natural Language Input (task instructions), Software Applications (automation targets), and Execution Environment (agent operational context), to evaluate their performance, reliability, and development effort.
- The study found that while RPA generally outperforms AACU in execution speed and reliability for repetitive, stable tasks, AACU significantly reduces development time and offers greater flexibility for dynamic interfaces.
- Despite current limitations in production readiness, AACU shows promise for rapid prototyping and lightweight automation, suggesting future research into hybrid RPA-AACU architectures and multi-agent orchestration.
MAGneT: Coordinated Multi-Agent Generation of Synthetic Multi-Turn Mental Health Counseling Sessions
- MAGneT (Coordinated Multi-Agent Generation of Synthetic Multi-Turn Mental Health Counseling Sessions): introduces a novel multi-agent framework for synthetic psychological counseling session generation that decomposes counselor response generation into coordinated sub-tasks, utilizing a Counselor Agent (generates counselor responses) with a CBT Agent (produces structured treatment plan), a Technique Agent (selects therapeutic techniques), five Specialized Response Agents (Reflection Agent (mirrors client expressions), Questioning Agent (explores client feelings), Solutions Agent (provides actionable solutions), Normalizing Agent (validates client experiences), Psycho-ed Agent (offers therapeutic information)), and a Response Generator (synthesizes final utterance), alongside a Client Agent (simulates client behavior) with an Intake Form (client profile, issues) and Attitudes (client emotional stance).
- This framework employs specialized LLM agents, each grounded in core therapeutic techniques, and coordinates them via a dynamic technique selector and a CBT-based planning agent to generate psychologically grounded and nuanced counseling dialogues.
- The system further simulates realistic client behavior using detailed profiles and attitude modeling, and integrates a unified evaluation framework for comprehensive assessment of generated counseling data quality and diversity.
TAGAL: Tabular Data Generation using Agentic LLM Methods
- TAGAL (Tabular Data Generation using Agentic LLM Methods): introduces a collection of training-free methods for generating synthetic tabular data using an agentic workflow, which includes a Generation LLM (generates tabular data), Feedback LLM (criticizes data, provides recommendations), Initial Prompt (guides initial data generation), Analysis Prompt (guides feedback LLM evaluation), Feedback Prompt (incorporates feedback for generation), Generated Data (synthetic tabular examples), Feedback (LLM-generated data critique/recommendations), Summary LLM (summarizes conversation history), and Refined Prompt (summarized prompt for generation), to iteratively refine generated data quality through feedback loops.
- The framework comprises three distinct methods—SynthLoop, ReducedLoop, and Prompt-Refine—each employing iterative feedback mechanisms to enhance data quality and utility, with Prompt-Refine specifically using a Summary LLM to create a Refined Prompt for efficient generation.
- It demonstrates performance comparable to state-of-the-art training-required models and often outperforms other training-free approaches, showcasing the potential of agentic LLM workflows for high-quality tabular data generation, even with limited original data.
Towards Stable and Personalised Profiles for Lexical Alignment in Spoken Human-Agent Dialogue
- Lexical Profile Construction and Evaluation: introduces a method for constructing stable, personalised lexical profiles from transcribed spoken data, leveraging POS categories and word-based n-grams, processed by SpaCy's nl_core_news_lg pipeline, and evaluated with recall, coverage, and cosine similarity metrics, to form a basis for lexical alignment in human-agent dialogue.
- The study determined that profiles built from 10 minutes of speech, including 5 items for adjectives and conjunctions and 10 items for adverbs, nouns, pronouns, and verbs each, offered the best balance between performance and data efficiency.
- These stable and representative lexical profiles are crucial for developing inclusive lexical alignment strategies in conversational agents, particularly for users with limited real-time input, such as individuals with dementia, by providing a robust basis for LLM-based response generation.
COT-SPACE: A THEORETICAL FRAMEWORK FOR INTERNAL SLOW-THINKING VIA REINFORCEMENT LEARNING
- CoT-Space (Chain-of-Thought Space): introduces a novel theoretical framework that recasts LLM reasoning from a discrete token-prediction task into an optimization process within a continuous, reasoning-level semantic space, including a policy model, reasoning-level states, a reasoning loss landscape, CoT length, solution minimums, noise scale, generalization error, and empirical loss.
- This framework models the LLM reasoning process as a trajectory towards a solution minimum in a reasoning loss landscape, providing a more intuitive and powerful lens for theoretical analysis of internal slow-thinking via RL.
- The analysis demonstrates that an optimal CoT length exists, balancing underfitting (due to insufficient reasoning depth) and overfitting (due to increased model complexity and noise sensitivity), analogous to an optimal learning rate in classical ML.
- Meta-Policy Reflexion (MPR): introduces a hybrid framework that consolidates LLM-generated reflections into a structured Meta-Policy Memory (MPM), which guides an LLM Base Policy through soft memory-conditioned decoding and hard admissibility checks (HAC) for training-free self-improvement.
- MPR externalizes reusable corrective knowledge as predicate-like rules, enforcing domain constraints to reduce unsafe actions without modifying LLM parameters, thereby retaining adaptability.
- The framework demonstrates consistent gains in execution accuracy and robustness on the AlfWorld benchmark compared to Reflexion baselines, with HAC further improving stability.
World Model Implanting for Test-time Adaptation of Embodied Agents
- WorMI (World Model Implanting): introduces a framework for embodied agents that combines LLMs' reasoning capabilities with independently learned, domain-specific world models through test-time composition, including a Reasoning Model (LLM), Pre-trained World Models, Prototype-based World Model Retrieval, Object Detection Model, Embedding Model, Prototypes, World-wise Compound Attention, Linear Projection Layer, World-level Cross-attention, Reasoning-level Cross-attention, and an Implant/Remove Mechanism.
- The framework's Prototype-based World Model Retrieval method efficiently selects relevant world models using trajectory-based abstract representation matching, while the World-wise Compound Attention method integrates and aligns knowledge from retrieved models with the reasoning model.
- This dual-stage design enables flexible, test-time fusion of domain-specific knowledge, enhancing adaptability to unseen domains and maintaining cross-domain adaptability for embodied agents.
VoxRole: A Comprehensive Benchmark for Evaluating Speech-Based Role-Playing Agents
- VoxRole: introduces a comprehensive benchmark for evaluating speech-based Role-Playing Conversational Agents (RPCAs), built using a Spoken Dialogue Extraction Pipeline (Extracts movie dialogues) and a Persona Distillation Pipeline (Builds character profiles), and evaluated with a multi-dimensional Evaluation Framework (Assesses model performance).
- The Spoken Dialogue Extraction Pipeline automatically extracts character-rich spoken dialogues from movies by aligning audio with scripts and curating semantically validated segments using components like FFmpeg, Resemble, Whisper-large-v3, Wav2Vec2.0, and MPNet.
- The Persona Distillation Pipeline leverages an LLM and an Acoustic Feature Extraction Module to systematically construct multi-dimensional character profiles, encompassing personality, linguistic style, relationships, and acoustic characteristics, which are then used to generate role-playing prompts for evaluation by an LLM Judge.
MobileRAG: Enhancing Mobile Agent with Retrieval-Augmented Generation
- MobileRAG (Retrieval-Augmented Generation): introduces a mobile agent framework, with InterRAG (external knowledge retrieval), LocalRAG (local app management), and MemRAG (historical operation memory), designed to enhance mobile agents for accurate user query identification and efficient complex mobile task execution.
- The framework addresses limitations of current LLM-based mobile agents, such as over-reliance on LLM comprehension, limited external interaction, and absence of effective memory.
- MobileRAG improves task completion, reduces operational steps, and enhances adaptability by leveraging external knowledge and learning from past successful operations.
FaMA: LLM-Empowered Agentic Assistant for Consumer-to-Consumer Marketplace
- FaMA (Facebook Marketplace Assistant): introduces an LLM-powered agentic assistant for C2C marketplaces, integrating a Llama 4 Maverick LLM as its core reasoning engine, a memory module (Scratchpad, Dialog History, Listings Information), and a suite of specialized Marketplace Tools (Listing Operation, Inventory Search, Messaging) along with RAG and a Knowledge Base.
- This conversational agent simplifies user experience by interpreting natural language commands to automate high-friction workflows for both buyers and sellers, including listing management, bulk messaging, and efficient product discovery.
- FaMA achieves a 98% task success rate and enables up to a 2x speedup in interaction time, providing a lightweight and accessible alternative to traditional app interfaces.
- MPDF (Meta-Policy Deliberation Framework): introduces a framework for multi-agent LLM collaboration, enabling agents to learn adaptive meta-cognitive policies through a Meta-Cognitive State Space (agent's internal cognitive status), Agent's Observation with Meta-Cognition (local and peer meta-cognitive states), a Policy Network (integrates self-assessment and social context), and a Deliberative Action Space (high-level strategic choices) optimized by SoftRankPO (stable policy optimization algorithm).
- This framework allows agents to dynamically adjust their behavior based on internal confidence and situational context, moving beyond fixed collaboration protocols to dynamic, deliberative strategies.
- SoftRankPO, a key component, stabilizes policy learning by converting raw rewards into rank-based advantages, ensuring robust convergence across diverse reward regimes.
Leveraging LLM-Based Agents for Intelligent Supply Chain Planning
- SCPA (Supply Chain Planning Agent): introduces an LLM-based agent framework for intelligent supply chain planning, featuring Input (user query), Memory (short-term/long-term storage), a Pre-trained LLM (foundation large language model), Task Management (planning orchestration), Task Execution (sub-task processing), and Output (planning report).
- The framework leverages LLM-based agents for intent classification, task orchestration, task execution, and iterative plan correction, enabling autonomous interpretation of natural language queries and dynamic adjustment of plans.
- SCPA demonstrates improved operational efficiency, plan accuracy, and stock availability in real-world e-commerce scenarios by providing evidence-based planning reports and automating complex supply chain decision-making.
SAMVAD: A Multi-Agent System for Simulating Judicial Deliberation Dynamics in India
- SAMVAD (Simulated Agent-based Multi-agent Verdict Adjudication): introduces a Multi-Agent System for simulating judicial deliberation in India, comprising an Orchestrator (manages simulation lifecycle), Judge Agent (generates impartial instructions), Prosecution Counsel Agent (constructs arguments for prosecution), Defense Counsel Agent (constructs arguments for defense), Adjudicator Agents (simulate judicial bench), LLMs (power agents' reasoning), RAG (grounds reasoning in knowledge), Legal Knowledge Base (authoritative Indian legal documents), Vector DB (stores legal document embeddings), Case Files (structured input for simulation), and Final Judgement (consensus-based verdict output).
- The system's core innovation is the deep integration of a domain-specific RAG pipeline, grounding LLM-powered agents in Indian legal texts to generate legally sound, citable instructions and arguments.
- This framework offers a configurable and explainable platform for exploring legal reasoning and group decision-making dynamics within the Indian judicial context, enhancing simulation fidelity and transparency.
- Mixed-Methods Framework for Stakeholder-Centric Summarization: introduces a systematic approach for generating tailored explanations of health simulations, with components including Simulation Model, Decomposition, RDF Representation, Finetuning, Visual Synthesis, LLM Prompting, Factuality Evaluation, Engagement Evaluation, Semi-structured Interviews, Summary Revision, Controllable Aspects, Validated Questionnaires, Quality Assessment, Preferred Aspects Identification, Contextual Information Retrieval, Optimization Strategies, Target Definition, LLM Optimization (Retriever, EAG-Sum, Direct Preference Optimization), Quantitative Evaluation, and Extrinsic Evaluation.
- This framework addresses the challenge of making complex health simulation models accessible to diverse stakeholders by tailoring LLM-generated summaries to their specific informational needs and stylistic preferences.
- It employs a two-step iterative process involving initial summary generation and evaluation, followed by optimization of LLMs and further assessment to ensure factual correctness and user engagement.
Maestro: Joint Graph & Config Optimization for Reliable AI Agents
- Maestro: introduces a framework-agnostic holistic optimizer for LLM agents that jointly searches over agent graphs and configurations to maximize agent quality, subject to explicit rollout/token budgets.
- It employs a block-coordinate scheme, alternating between C-step (configuration updates) and G-step (graph updates), guided by both numeric and reflective textual feedback from execution traces.
- The framework supports flexible agent-graph spaces, including branching, memory/state nodes, tool-augmented subroutines, and multi-model/multi-tool choices with tunable hyperparameters, addressing structural failure modes that prompt tuning alone cannot fix.
- LLM-Automated Design and Plasticity-Based Evaluation Framework: introduces a closed-loop system for training neural organoids in scalable virtual environments, leveraging an LLM for automated experimental protocol design and evaluating learning through synaptic plasticity.
- The framework includes three distinct virtual environments (conditional avoidance, predator-prey, Pong) with increasing complexity, formalizing state/action spaces, sensory encoding, motor decoding, and feedback protocols.
- The LLM acts as a meta-controller, generating and optimizing experimental protocols, and is refined through prompt engineering and fine-tuning based on collected performance and neurophysiological data.
Bootstrapping Task Spaces for Self-Improvement
- EXIT (Exploratory Iteration): introduces an autocurriculum RL method for LLMs to perform multi-step self-improvement at inference-time, training on informative single-step iterations by growing a task space.
- The framework leverages an LLM fine-tuned with GRPO, dynamically sampling self-iteration task instances from a Task Buffer, guided by a Learning Potential Score, and enhanced by Self-Divergence Steps and a Diversity Bonus in an Embedding Space.
- This approach enables robust self-improvement and generalization across various domains, including competition math, multi-turn tool-use, and ML engineering, by efficiently generating and learning from diverse task instances within a Search Scaffold.
EMERGENT SOCIAL DYNAMICS OF LLM AGENTS IN THE EL FAROL BAR PROBLEM
- LLM Agents: introduces a simulation of LLM agents in a spatially extended El Farol Bar problem, where each autonomous decision-maker, powered by an LLM (GPT-4o), generates messages, memories, and actions within a 2D grid environment, influenced by a crowding threshold and communication radius, guided by a prompt.
- The simulation reveals emergent social dynamics, including spontaneous motivation, collective decision-making, and the development of human-like bounded rationality and differentiated social roles among agents.
- These findings demonstrate how LLM agents balance game-theoretic rationality with culturally-encoded social motivations, providing a new paradigm for studying complex social systems.
Narrative-to-Scene Generation: An LLM-Driven Pipeline for 2D Game Environments
- Narrative-to-Scene Generation Pipeline: introduces a lightweight pipeline that transforms short narrative prompts into a sequence of 2D tile-based game scenes, reflecting the temporal structure of stories, with LLM (story generation, predicate extraction, terrain suggestion), Narrative Parsing (story summarization, time frame segmentation, predicate triples extraction), Semantic Matching (object-tile alignment, affordance filtering), Scene Synthesis (procedural terrain generation, spatial object placement, layered 2D scene rendering), GameTileNet Dataset (visual asset repository, semantic embeddings), and Knowledge Graph Module (symbolic reasoning, temporal integration).
- The pipeline segments LLM-generated narratives into three key time frames, extracts "Object-Relation-Object" predicate triples, and retrieves visual assets using affordance-aware semantic embeddings from the GameTileNet dataset.
- It generates layered terrain using Cellular Automata and places objects using spatial rules grounded in the predicate structure, ensuring semantic coherence and narrative alignment in 2D game environments.
Bootstrapping Reinforcement Learning with Sub-optimal Policies for Autonomous Driving
- Bootstrapping RL via Suboptimal Policy Framework: introduces a novel DRL-based autonomous driving framework that integrates a suboptimal controller to guide RL agents, enhancing exploration and learning efficiency in complex driving scenarios.
- The framework leverages the suboptimal policy both as a soft constraint on the RL policy during initial training and as a source for populating the replay buffer with additional training samples.
- This approach enables the RL agent to overcome exploration barriers and converge on optimal driving policies by providing plausible and human-like behavior.
In-Context Policy Adaptation via Cross-Domain Skill Diffusion
- ICPAD (In-Context Policy Adaptation): introduces a framework for rapid policy adaptation in long-horizon multi-task environments, leveraging cross-domain skill diffusion and dynamic domain prompting to adapt skill-based RL policies to diverse target domains with limited data.
- The framework learns domain-agnostic prototype skills and a domain-grounded skill adapter jointly from offline data through cross-domain consistent diffusion processes, which are then adapted in-context using dynamic domain prompting.
- This approach facilitates unified policy adaptation by using prototype skills as a middle-tier layer, translating them into domain-specific actions via dedicated skill adapters guided by retrieval-based attention from few-shot target data.
SasAgent: Multi-Agent AI System for Small-Angle Scattering Data Analysis
- SasAgent (Multi-Agent AI System for Small-Angle Scattering Data Analysis): introduces a multi-agent AI system for automating small-angle scattering (SAS) data analysis, featuring a Coordinator Agent (interprets, delegates), Generation Agent (generates synthetic data), Fitting Agent (fits experimental data), SLD Agent (calculates SLD), Model Data Tool (generates scattering function), RAG Documentation Tool (provides SasView documentation), Bump Fitting Tool (executes SasView fitting), SLD Calculator Tool (calculates SLD), and a Gradio-based User Interface (Gradio-based), all powered by LLMs (power agents).
- The system is self-aware, capable of performing SLD calculation, synthetic data generation, and experimental data analysis, while guiding users through intuitive text prompts and data uploads.
- Implemented using CrewAI and a Gradio-based web interface, SasAgent leverages LLM-friendly tools derived from the SasView Python library to streamline scientific workflows and enhance automation in SAS research.
SYNTHESIZING SHEET MUSIC PROBLEMS FOR EVALUATION AND REINFORCEMENT LEARNING
- Data Synthesis Framework: introduces a novel method for generating verifiable sheet music problems grounded in music theory, serving as both evaluation benchmarks and training data for reinforcement learning with verifiable rewards (RLVR).
- This framework programmatically generates sheet music questions and answers in both textual (ABC notation) and visual (staff notation image) modalities, without reliance on LLMs.
- The resulting Synthetic Sheet Music Reasoning Benchmark (SSMR-Bench) and training set enhance LLMs' and MLLMs' reasoning abilities in sheet music understanding and facilitate AI-assisted music creation.
DEPTH-BREADTH SYNERGY IN RLVR: UNLOCKING LLM REASONING GAINS WITH ADAPTIVE EXPLORATION
- DARS (Difficulty Adaptive Rollout Sampling): introduces a framework to unlock LLM reasoning gains by addressing depth and breadth dimensions in RLVR, utilizing pre-rollout difficulty estimation and multi-stage rollout re-balancing to re-weight hard problems.
- The framework includes two schedules, Equal-Treatment (ET) and Hardness-Weighted (HW), for rebalancing cumulative advantage, and can be augmented with Large Breadth Training (DARS-Breadth) by replacing PPO mini-batch updates with full-batch updates for synergistic performance.
- DARS improves Pass@K performance by focusing on hard problems, while DARS-Breadth further enhances Pass@1 by sustaining exploration and reducing gradient noise through increased training data breadth.
- LLM-based Agents: introduces a study on how delegating identity-defining choices to LLM-based agents, including Generic AI Agents, Personalized AI Agents with a User Profile Generator and User Data Input, and a Core LLM, impacts interpersonal distinctiveness and intrapersonal diversity of people's choices.
- The research compares choices made by generic and personalized LLM agents against a human baseline, using real-world Facebook page preferences from 1,000 users to measure distinctiveness and diversity.
- Findings indicate that both agent types reduce choice distinctiveness, with personalized agents more strongly compressing intrapersonal diversity, highlighting a trade-off between distinctiveness and diversity in AI-assisted decision-making.
REAL-TIME INSTRUMENT PLANNING AND PERCEPTION FOR NOVEL MEASUREMENTS OF DYNAMIC PHENOMENA
- Dynamic Plume Planning: introduces an automated workflow for real-time instrument planning and perception, synthesizing look-ahead satellite imagery acquisition, onboard data analysis, plume classification, denoising, and autonomous trajectory planning to obtain pinpoint measurements of dynamic phenomena like volcanic plumes.
- The workflow leverages computer vision and machine learning classifiers, including U-Net architectures, for plume segmentation, followed by morphological operations for denoising, and employs various trajectory planning algorithms to guide a Narrow Field of View (NFOV) sensor.
- This onboard system significantly increases the science utility return of high-resolution instruments by dynamically targeting transient events, demonstrating efficient runtimes and generalizability to other remote sensing applications.
- A2AI (Aspective Agentic AI): introduces a bottom-up framework that situates AI agents in their environment, with all behaviors triggered by changes in their environments, including Environment (central data store), Aspects (specialized environment views), p-agent (generative, perceptive agent) (generates, perceives aspects), a-agent (action agent) (requests environment changes), Agent (operates within aspect), Change Request (action agent's modification proposal), Change List (environment modification summary), Change Summary (perceptive agent's aspect update), and Human (initiates change requests).
- This framework enables selective information disclosure by allowing agents to perceive only limited "aspects" of their environment, preventing information leakage and enhancing security and computational efficiency.
- The reactive, asynchronous, and bottom-up architecture, inspired by situated AI, ensures agents dynamically respond to environmental changes while maintaining strict information isolation.
- GCSL-NF (Goal-conditioned Supervised Learning with Negative Feedback): introduces a novel model that integrates contrastive learning principles into the GCSL framework to learn from both success and failure, including a Policy (πθ), Qθ (Q-function), Replay Buffer (R), Relabelled Goal Trajectory Dataset (T+), Original Goal Trajectory Dataset (To), Imitation Learning Loss (L+), Negative Feedback Loss (Lo), Similarity Function (pφ), and Combined Loss.
- This approach addresses limitations of GCSL by utilizing both relabelled successful experience and failures, enabling agents to learn from mistakes and overcome inherent biases.
- The framework employs a learned distance function to assess the quality of achieved states relative to intended goals, promoting exploration and avoiding behavioral stagnation.
Towards Agentic OS: An LLM Agent Framework for Linux Schedulers
- SchedCP (LLM Agent Framework for Linux Schedulers): introduces a decoupled control plane and multi-agent LLM system to autonomously optimize Linux schedulers, featuring a Model Context Protocol server, Workload Analysis Engine, Scheduler Policy Repository, Execution Verifier, and sched-agent's Observation, Planning, Execution, and Learning Agents.
- This framework separates AI's semantic reasoning from the system's execution, enabling LLM agents to safely and efficiently generate and deploy custom eBPF scheduling policies without human intervention.
- The framework achieves significant performance improvements and cost reductions by bridging the semantic gap between application needs and kernel policies through iterative refinement and continuous learning.
Language Models Do Not Follow Occam's Razor: A Benchmark for Inductive and Abductive Reasoning
- INABHYD (Inductive and Abductive Hypothesis Discovery) introduces a programmable and synthetic dataset for evaluating LLMs' inductive and abductive reasoning capabilities, comprising reasoning examples with an incomplete world model, observations, and hypotheses evaluated by an Occam's Razor-based metric.
- The dataset challenges LLMs to generate high-quality hypotheses to explain observations under an incomplete world model, structured as an ontology tree, with difficulty characterized by the tree's height.
- The research reveals that LLMs struggle with complex world models and producing high-quality hypotheses, even with reasoning-enhancing techniques, highlighting limitations in non-deductive reasoning.
- EvolveSignal: introduces an LLM-powered coding agent that iteratively refines an Initial Program (starting algorithm) into a Discovered Program (optimized algorithm) by using a Program Database (stores programs/metrics), Prompt Sampler (constructs LLM prompts), LLMs Ensemble (generates code modifications), and Evaluators Pool (simulates/scores programs) to evaluate Child Programs (modified algorithm).
- The framework formulates fixed-time signal control as a program synthesis problem, where LLMs generate Python functions representing algorithms, which are then optimized through simulation-based evaluation and evolutionary search.
- Experiments demonstrate that the discovered algorithms outperform baselines in delay reduction and stop minimization, providing interpretable modifications and practical insights for traffic engineers.
- VulnRepairEval: introduces an exploit-based evaluation framework for assessing LLM vulnerability repair capabilities, featuring Patch Generation, Runtime Injection, Automatic Deployment, Container Execution, and Result Analysis modules, designed for reproducible differential assessment.
- The framework leverages functional Proof-of-Concept (PoC) exploits to verify patch success, requiring the original exploit to fail against the modified code in a containerized environment.
- This work reveals that current LLMs struggle with precise vulnerability localization and syntactically/logically correct patch generation, with advanced prompting and multi-agent approaches yielding minimal improvements.
AGENTRACER: WHO IS INDUCING FAILURE IN THE LLM AGENTIC SYSTEMS?
- AgenTracer: introduces an automated framework for annotating failed multi-agent trajectories and a lightweight failure tracer, AgenTracer-8B, which leverages counterfactual replay, programmatic fault injection, and multi-granular reinforcement learning to efficiently diagnose errors in LLM agentic systems.
- The framework generates a curated dataset, TracerTraj, of over 2,000 high-fidelity failure trajectories, enabling the training of AgenTracer-8B to achieve state-of-the-art performance in agentic system failure attribution.
- AgenTracer-8B provides actionable feedback to off-the-shelf multi-agent systems, leading to performance gains and empowering self-correcting and self-evolving agentic AI.
AIVA: An AI-based Virtual Companion for Emotion-aware Interaction
- AIVA (AI-based Virtual Companion for Emotion-aware Interaction): introduces an AI-based virtual companion that integrates multimodal sentiment perception into LLMs, enabling emotionally aligned and animated Human-Computer Interaction (HCI) through its Multimodal Sentiment Perception Network (MSPN), Vision Transformer (ViT), Textual Encoder (BERT), Cross-Attention (CA) mechanism, Cross-Modal Fusion Transformer, Sentiment Prototypes, Classifier (MLP), Large Language Model (LLM), Emotion-aware Prompt Engineering (EPE), Text-to-Speech (TTS) system, and Animated Avatar module (Live2D).
- The framework's MSPN component processes multimodal inputs (language, facial expressions, voice) to extract sentiment signals, which are then injected into the LLM via EPE to generate contextually appropriate and empathetic language responses.
- AIVA further enhances user experience by providing expressive verbal and visual feedback through its TTS system and an animated avatar module, creating natural, engaging, and emotionally aligned interactions for applications in companion robotics, social care, and mental health.
Loong: Synthesize Long Chain-of-Thoughts at Scale through Verifiers
- Loong Project: introduces an open-source framework for scalable synthetic data generation and verification, featuring LOONGBENCH (curated seed dataset) and LOONGENV (synthetic data generation environment) with a Generator, Environment, Trainable Agent, and Verifier.
- The framework establishes an agent-environment loop where an LLM-based Generator creates synthetic questions and executable code, which the Environment runs to produce answers, then a Trainable Agent generates Chain-of-Thought solutions, and a Verifier compares these for an RL reward.
- This system aims to overcome the scarcity of high-quality, verifiable datasets in reasoning-intensive domains beyond mathematics and programming, enabling large-scale reinforcement learning with minimal human supervision.
- DiaCBT (Long-Periodic Dialogue Corpus Guided by Cognitive Conceptualization Diagram for CBT-based Psychological Counseling): introduces a long-periodic dialogue corpus for CBT-based psychological counseling, with Case Annotation, Cognitive Conceptualization, CCD-guided Dialogue Generation, Expert Evaluation, LLMs, CBT Segments, Cognitive Conceptualization Diagrams (CCDs), Human Annotators, Experts, Client Simulator (GPT-4o-mini), Grader Model (LLMrwd), and Therapist Model (fine-tuned LLM therapist), where the paper constructs a multi-session dialogue corpus guided by CCDs to enhance LLMs' ability to emulate CBT psychologists.
- The framework leverages LLMs to generate CCDs for diverse client scenarios and then uses these CCDs, along with annotated CBT strategies, to create realistic, multi-session counseling dialogues.
- DiaCBT also includes a comprehensive evaluation framework, employing a client simulator and a grader model to benchmark the performance of LLM-based therapists against established psychological criteria.
InstaDA: Augmenting Instance Segmentation Data with Dual-Agent System
- InstaDA (Augmenting Instance Segmentation Data with Dual-Agent System): introduces a novel dual-agent system for instance segmentation data augmentation, featuring a T-Agent (generates diverse synthetic data) and an I-Agent (augments data from training images), along with a Prompt Rethink mechanism, BiRefNet, CLIP dual-similarity, Soft-Edge Maps Fusion, ControlNet, Image2Image, SAM-box, and Copy-Paste.
- The T-Agent leverages LLMs and diffusion models with a Prompt Rethink mechanism to iteratively refine prompts and generate diverse images, while the I-Agent enriches data distribution by generating new instances conditioned on existing training images using ControlNet and Image2Image.
- The framework ensures high-quality annotations through BiRefNet and SAM-box for segmentation, and filters generated instances using CLIP dual-similarity and CLIP score before integrating them via Copy-Paste to enhance dataset diversity and distribution.
- app.build (Environment Scaffolding): introduces a production framework for scaling agentic prompt-to-app generation, which wraps LLMs with an Orchestrator, stack-specific Actors, Sandbox Manager, Validation Layer, and Task Runner to provide systematic validation and structured environments for reliable application development.
- The framework combines multi-layered validation pipelines, stack-specific orchestration, and a model-agnostic architecture, implemented across three reference stacks (TypeScript/tRPC, PHP/Laravel, Python/NiceGUI).
- Through evaluation on 30 generation tasks, the framework achieves a 73.3% viability rate with 30% perfect quality scores, demonstrating that scaling reliable AI agents requires scaling environments, not just models.
CoreThink: A Symbolic Reasoning Layer to reason over Long Horizon Tasks with LLMs
- CoreThink (General Symbolics Reasoning): introduces a state-of-the-art reasoning layer built upon a novel reasoning method called General Symbolics, with Native Language Parsing & Semantic Preservation, In-Language Reasoning Architecture, Execution & Explainability, Representational Translation Avoidance, Computational Optimization Layer, Agentic Coding IDE, and ARC-AGI-2 Neuro-Symbolic Pipeline, where it provides a pure performance uplift for LLMs on long-horizon reasoning tasks without fine-tuning or training costs.
- The framework operates on a pure natural language-to-natural language basis, avoiding representational loss and brittleness associated with translating human language into formal logic or high-dimensional vectors.
- It achieves state-of-the-art performance across tool-calling, code generation, and planning benchmarks, demonstrating robust capabilities in complex, multi-step algorithmic reasoning and software engineering challenges.
L-MARS: Legal Multi-Agent Workflow with Orchestrated Reasoning and Agentic Search
- L-MARS (Legal Multi-Agent Workflow with Orchestrated Reasoning and Agentic Search): introduces a multi-agent workflow that integrates structured reasoning, agentic search, and sufficiency verification for legal question answering, including a Query Agent (parses, refines queries), Search Agent (executes retrieval tools), Serper (online web search), Local RAG (offline document retrieval), BM25 (ranking function), CourtListener (case law API), Judge Agent (verifies evidence, refines), Summary Agent (synthesizes final answer), WorkflowState (manages system state), and LangGraph (orchestrates workflow), to reduce hallucination and uncertainty in legal QA.
- The system operates in two modes: Simple Mode, a single-pass retrieval-summarization pipeline, and Multi-Turn Mode, which adds Judge Agent-guided iterations with sufficiency checks and query refinement for enhanced accuracy.
- This iterative reasoning-search-verification loop maintains coherence, filters noisy evidence, and grounds answers in authoritative law, demonstrating a scalable and reproducible blueprint for deploying LLMs in high-stakes legal domains.
An Agentic Model Context Protocol Framework for Medical Concept Standardization
- MCP (Model Context Protocol): introduces a zero-training, hallucination-preventive mapping system for medical concept standardization, featuring an Input (user query), an Agentic LLM (interprets, reasons, calls tools), MCP (standardized, secure framework), MCP Resources (contextual guidance, preferences), Reasoning 1: keyword inference (interprets user input), Athena OHDSI API (external vocabulary service), Concept list (candidate medical concepts), Reasoning 2: concept selection (selects best concept), and Output (standardized OMOP concept), which enables explainable mapping and improves efficiency and accuracy in mapping source medical terms to OMOP standard concepts.
- The system leverages LLMs with real-time access to external resources like OHDSI Athena, guided by OMOP data model specifications, documentation, and vocabulary preferences, to mitigate hallucination and ensure clinically appropriate mappings.
- This framework provides a robust, auditable, and user-guided solution for medical terminology mapping, suitable for both exploratory and production environments without requiring fine-tuning or complex infrastructure.
What Would an LLM Do? Evaluating Policymaking Capabilities of Large Language Models
- LLM-ABM Integration Pipeline: introduces a novel benchmark and an automated pipeline to evaluate LLMs' policymaking capabilities for homelessness alleviation, with all its components, where the framework assesses LLM alignment with human experts and simulates policy impacts.
- This framework assesses LLM alignment with human experts on policy choices across four geographies, grounding policies in the Capability Approach for human development.
- The pipeline connects LLM-generated policy proposals to an agent-based model to explore their social impact through simulated scenarios, offering insights into scalable and non-invasive social policymaking.
- Gaze Analytics Dashboard with Conversational AI Support: introduces a user-centered dashboard for English Language Arts (ELA) instruction, integrating gaze heatmaps, student performance tables, score trajectories, scanpaths, and an LLM-powered conversational agent for interpreting multimodal learning analytics.
- The system leverages eye-tracking technology to capture student gaze data, which is then visualized and summarized through an LLM-Augmented Report Generation Pipeline to provide actionable insights for teachers and students.
- This iterative design, guided by user feedback and data storytelling principles, aims to make complex gaze analytics approachable and pedagogically valuable, enhancing instructional decision-making and student reflection.
ARE LLM AGENTS BEHAVIORALLY COHERENT? LATENT PROFILES FOR SOCIAL SIMULATION
- Framework for Probing Behavioral Coherence: introduces a study to evaluate the internal consistency of LLM agents by eliciting their internal states (preference and openness) and observing their behavior in dialogue settings, using an LLM-as-judge to score agreement.
- The study reveals significant internal inconsistencies in LLMs across various models and sizes, showing that agents often suppress disagreement and favor positive sentiment, even when explicitly biased.
- These findings highlight a critical gap in LLM capabilities, as agents fail to maintain behavioral coherence over time, questioning their reliability as substitutes for human participants in social science research.
Learning When to Plan: Efficiently Allocating Test-Time Compute for LLM Agents
- Dynamic Planning Agent Architecture: introduces a framework for LLM agents to dynamically allocate test-time compute for planning, with LLM Agent (unified output generator), Context (agent's internal state), New Observation (current environment state), Previous Plan (prior strategic guidance), Implicit Plan Decision (decides planning necessity), Planning Policy (generates new strategic plan), Acting Policy (executes next action), Generated Output (action or plan and action), Action (command to environment), New Plan (updated strategic guidance), and Environment (sequential decision-making tasks).
- This architecture enables a single, monolithic LLM to implicitly decide when to plan by generating a
<plan>token, then parsing its unified output to extract both the action and, if present, the new plan. - The framework is trained using a two-stage pipeline of supervised fine-tuning and reinforcement learning, allowing agents to learn strategic planning, plan execution, and replanning only when necessary, optimizing computational resource allocation.
- SHT-driven Framework (Sequential Hypothesis Testing-driven Framework): introduces a novel approach for adversarial decision-making in partially observable multi-agent systems, modeling deception as a dynamic optimization problem within a leader-follower Stackelberg game.
- This framework integrates a Blue Team Agent (follower) that uses a Linear-Quadratic Control Framework and an Optimal Control Solution Module for strategic misdirection, and a Red Team Agent (leader) that employs a Sequential Hypothesis Testing (SHT) Module and a Red Team Strategy Optimization Module (using FPI, NN, or FBS algorithms) for counter-deception.
- The system utilizes Stochastic Differential Equations (SDEs) for state dynamics, Cost Functionals for objective quantification, a Likelihood Ratio Statistic (LT) for deception effectiveness, and a Regularization Penalty Term to model skepticism, providing insights into strategic deception and counter-deception.
AutoGrid AI: Deep Reinforcement Learning Framework for Autonomous Microgrid Management
- AutoGrid AI (Deep Reinforcement Learning Framework for Autonomous Microgrid Management): introduces a deep reinforcement learning framework for autonomous microgrid management, integrating transformer architecture for forecasting renewable generation and a PPO agent for decision-making in a simulated environment.
- The framework optimizes microgrid energy dispatch strategies to minimize costs and maximize renewable energy utilization, demonstrating improvements in energy efficiency and operational resilience compared to rule-based methods.
- It also provides an open-source framework for simulating various microgrid environments, supporting the development of zero-carbon energy systems.
Advancing SLM Tool-Use Capability using Reinforcement Learning
- GRPO (Group Relative Policy Optimization): introduces a novel reward model, optimized with GRPO, for fine-tuning Small Language Models (SLMs) to master structured tool use, ensuring valid JSON output, precise tool selection, and accurate parameter specification, utilizing components such as Extraneous Text Penalty, JSON Validity Reward, Function Name Reward, Argument Matching, Penalty for Extra Tool Calls, and Capability-Aware Reward Modeling.
- The approach employs a strict zero-reward mechanism for extraneous text and over-generation of tool calls, alongside a capability-aware reward modeling strategy that iteratively refines the reward function based on observed learning behavior.
- This method significantly boosts SLM tool-use accuracy (6x-21x improvements) and computational efficiency, making tool-augmented AI agents more deployable in resource-constrained environments.
BEYOND CORRECTNESS: HARMONIZING PROCESS AND OUTCOME REWARDS THROUGH RL TRAINING
- PROF (PRocess consistency Filter): introduces a data curation strategy that harmonizes noisy, fine-grained Process Reward Models (PRMs) with accurate, coarse-grained Outcome Reward Models (ORMs) through consistency-driven sample selection, which includes initial rollouts, PRM, ORM, consistency score, correct group, incorrect group, sample selection/filtering, policy update, and Group Relative Policy Optimization (GRPO).
- The framework over-samples responses, then ranks and filters them by consistency between PRMs and ORMs, removing samples where process and outcome signals conflict to eliminate conflicting and noisy gradients.
- PROF improves final accuracy and intermediate reasoning step quality by retaining correct responses with higher averaged process values and incorrect responses with lower averaged process values, while maintaining a balanced training sample ratio.
AppCopilot: Toward General, Accurate, Long-Horizon, and Efficient Mobile Agent
- AppCopilot: introduces a multimodal, multi-agent, general-purpose on-device assistant, with Multimodal Foundation Models (core for perception, reasoning, action), OCR+OR Module (identifies UI elements, bounding boxes), Multi-Agent Collaborative Decision-Making Strategy (aggregates actions from multiple agents), Reinforcement Learning (optimizes long-horizon task policies), High-level Planning Agent (decomposes tasks, allocates resources), Personalized Information Memory and Retrieval Mechanism (stores and retrieves user preferences), Experience Reuse Framework (replays historical successful tasks), and Hybrid Control Framework (integrates GUI and API control), which operationalizes an end-to-end autonomous pipeline for mobile agents from data to deployment, addressing generalization, accuracy, long-horizon capability, and efficiency.
- The system integrates multimodal foundation models for robust Chinese-English support, combining chain-of-thought reasoning, hierarchical task planning, and multi-agent collaboration at the reasoning and control layer.
- At the execution layer, it enables user personalization, voice interaction, function/tool calling, cross-app and cross-device orchestration, and comprehensive mobile app support, incorporating profiling-driven optimization for latency, memory, and energy.
Do LLM Modules Generalize? A Study on Motion Generation for Autonomous Driving
- LLM-to-Driving Transfer Analysis: introduces a systematic evaluation of five key LLM modules—tokenizer design, positional embedding, pre-training paradigms, post-training strategies, and test-time computation—within the context of motion generation for autonomous driving.
- The study demonstrates that, when appropriately adapted, these modules can significantly improve performance for autonomous driving motion generation, identifying effective transfer techniques and necessary domain-specific adaptations.
- The research provides insights into the transferability of LLM modules, validating their effectiveness on the Waymo Sim Agents benchmark and achieving competitive results.
- EDSAC (Ethical Decision Support Assessment Checklist): introduces a novel framework for evaluating LLMs' ethical performance in construction project management, assessing responses across seven critical dimensions: Ethical Soundness, Legal Compliance, Fairness/Non-Bias, Transparency/Explainability, Contextual Relevance, Practical Actionability, and Bias Sensitivity.
- The study employs a mixed-methods design, combining quantitative performance testing of LLMs against real-world ethical scenarios using EDSAC with qualitative analysis from semi-structured interviews with industry experts.
- Findings reveal LLMs demonstrate deficiencies in contextual nuance, accountability, and transparent reasoning, advocating for robust human-in-the-loop oversight and positioning LLMs as decision-support aids rather than autonomous ethical agents.
The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
- Agentic RL: introduces a paradigm for LLMs, reframing them as autonomous decision-making agents with Planning, Tool Use, Memory, Self-Improvement, Reasoning, and Perception modules, enabling complex, dynamic interactions.
- This approach contrasts with traditional LLM-RL by optimizing semantic-level behaviors in variable, partially observable environments through sequential decision-making loops.
- Reinforcement learning serves as the critical mechanism to transform these core capabilities from static, heuristic modules into adaptive, robust agentic behavior for general-purpose AI.
UI-TARS-2 Technical Report: Advancing GUI Agent with Multi-Turn Reinforcement Learning
- UI-TARS-2: introduces a native GUI-centered agent model, with a data flywheel (systematic training methodology), an all-in-one GUI sandbox (unified sandbox platform), a multi-turn reinforcement learning framework (stabilized RL framework), and parameter interpolation (merging specialized agents), designed to handle structured computer-use tasks and dynamic, game-like interactive environments.
- The framework employs a ReAct paradigm for agent formulation, interleaving reasoning, action, and observation, supported by a hierarchical memory state for context preservation and an interactive annotation platform for human-in-the-loop data generation.
- Its multi-turn RL framework incorporates advanced techniques like asynchronous agent rollout, stateful environment integration, reward shaping, and value pretraining to ensure stable and efficient learning across diverse, long-horizon tasks.
- FlexNGIA 2.0: introduces an Agentic AI-driven Internet architecture that leverages LLM-based AI agents, encompassing agents for application analysis, information fusion, SFC and protocol design, congestion control, resource allocation, monitoring, and failure management, to autonomously orchestrate, configure, and evolve the network.
- Each LLM-based agent is equipped with a Brain (LLM for reasoning), Memory (persistent context store), Planning (task decomposition), and Tools (external system interface), enabling autonomous, context-aware decision-making and real-time adaptation to dynamic network conditions and application demands.
- The framework redefines network architecture by embedding cognitive intelligence, allowing AI agents to redesign network protocols, logic, and algorithms on the fly, thereby delivering flexibility, intelligence, and responsiveness for diverse and evolving application requirements.
- SoD-RL Zonal Dispatching (Semi-on-Demand Transit Feeder Service with Reinforcement Learning-Based Zonal Dispatching Control): introduces a novel transit feeder service, with RL Model (decision-making agent) comprising Policy (Actor Model) (proposes actions) and Value Function (Critic Model) (evaluates actions), interacting with FleetPy Simulation (Environment) (simulates transit system) which includes Fleet Control (manages vehicle operations), Network (provides travel data), and Trip offer (processes passenger requests), through State (system observation), Action (dispatching decision), and Reward (performance feedback).
- This framework dynamically assigns Shared Autonomous Vehicles (SAVs) to subdivided flexible-route zones using a deep RL policy gradient algorithm (Proximal Policy Optimization) to respond to real-time demand fluctuations and operational needs.
- The system aims to maximize passengers served while maintaining frequent service on fixed-route portions, demonstrating improved efficiency and passenger service compared to traditional fixed-route and nominal semi-on-demand services.
Safety-Critical Multi-Agent MCTS for Mixed Traffic Coordination at Unsignalized Roundabout
- SC-MCTS (Safety-Critical Multi-Agent Monte Carlo Tree Search): introduces a safety-critical decision-making framework for autonomous vehicles navigating unsignalized, dual-lane roundabouts, with Problem Formulation as MDP, Safety-Critical Decision Making, Multi-Agent MCTS, and Reward Function Design and Optimization components, enabling cooperative decision-making by integrating deterministic and probabilistic prediction models.
- The framework employs a hierarchical safety assessment module to address AV-to-AV, AV-to-HDV, and AV-to-Road interactions through dynamic safety thresholds and spatiotemporal risk evaluation.
- An adaptive HDV behavior prediction scheme, combining the Intelligent Driver Model with probabilistic uncertainty modeling, and a multi-objective reward optimization strategy jointly considering safety, efficiency, and cooperative intent, further enhance the system's robustness in mixed traffic.
Plan Verification for LLM-Based Embodied Task Completion Agents
- Plan Verification Framework: introduces an iterative verification framework for LLM-based embodied task completion agents, featuring a Planning Agent (generates and revises plans), a Judge LLM (critiques action sequences), and an Iterative Refinement Loop (manages repeated critique and revision).
- This framework enables the Planning Agent to generate candidate plans and the Judge LLM to analyze and flag erroneous actions, such as redundant, contradictory, or missing steps, with natural language explanations.
- The iterative process refines action sequences, leading to progressively cleaner and more spatially coherent trajectories, thereby providing higher-quality training data for imitation learning in embodied AI.
Deep Research is the New Analytics System: Towards Building the Runtime for AI-Driven Analytics
- Our Prototype (extending Palimpzest): introduces a new runtime for AI-driven analytics, combining Deep Research flexibility with optimized semantic operator execution through
PZ.Context,Search Operator,Compute Operator,CodeAgents,Semantic Operators,ContextManager, andToolscomponents. - The prototype leverages
CodeAgents to physically implementSearchandComputeoperators, enabling dynamic planning and execution of optimized semantic operator programs over large unstructured datasets. - By introducing the
PZ.Contextabstraction with indexing and user-definedTools, and aContextManagerfor caching, the system aims to improve query performance and reduce computational costs for AI-driven analytics.
Contemporary Agent Technology: LLM-Driven Advancements vs Classic Multi-Agent Systems
- LLM-based Agent Architecture: introduces a comprehensive reflection on contemporary agent technology, contrasting LLM-driven advancements with classic Multi-Agent Systems by detailing the architectural pillars that define these new systems.
- The paper critically analyzes how recent LLM developments relate to foundational Multi-Agent Systems (MAS) concepts, models, and characteristics, emphasizing the shift from symbolic to sub-symbolic AI.
- It identifies key challenges and promising future directions in this rapidly evolving domain, highlighting the need for standardization and robust hybrid systems that combine formal principles with adaptive reasoning.
GridMind: LLMs-Powered Agents for Power System Analysis and Operations
- GridMind: introduces a multi-agent AI system that integrates LLMs with deterministic engineering solvers to enable conversational scientific computing for power system analysis, with a Planner/Coordinator (orchestrates agents, workflows), ACOPF Agent (handles AC Optimal Power Flow), CA Agent (performs T-1 reliability assessment), LLM (provides core reasoning), Memory (maintains analytical coherence), Tools (invokes deterministic solvers), Grid Data (provides power system information), Secure Access Data (ensures secure data retrieval), and Conversational Interface (manages user interaction).
- The system employs specialized agents coordinating AC Optimal Power Flow and N-1 contingency analysis through natural language interfaces, maintaining numerical precision via function calls to external tools and solvers.
- GridMind addresses workflow integration, knowledge accessibility, context preservation, and expert decision-support augmentation, demonstrating how conversational interfaces can enhance accessibility while preserving numerical rigor for critical engineering applications.
KUBEINTELLECT: A MODULAR LLM-ORCHESTRATED AGENT FRAMEWORK FOR END-TO-END KUBERNETES MANAGEMENT
- KubeIntellect: introduces a modular LLM-orchestrated agent framework for end-to-end Kubernetes management, featuring a User Interaction Layer, Query Processing Module, Task Orchestration Module with Memory, an Agent & Tool Execution Layer with specialized agents (including a Code Generator Agent), a Kubernetes Interaction Layer, and a Supporting System with an LLM Gateway, Persistent Context Service, and Security & Governance.
- This framework enables natural language interaction for comprehensive Kubernetes API operations, supporting dynamic tool synthesis, structured workflows, human-in-the-loop clarification, and secure execution across diverse workloads.
- KubeIntellect integrates memory checkpoints and a LangGraph-based orchestration engine, achieving a 93% tool synthesis success rate and 100% reliability in managing complex Kubernetes infrastructure.
- BioBlue: introduces a set of benchmarks for evaluating LLMs in long-running scenarios, including Sustainability (resource balance evaluation), Single-objective homeostasis (single metric stability), Multi-objective homeostasis (multiple metric stability), and Balancing unbounded objectives with diminishing returns (multi-goal optimization).
- These benchmarks reveal systematic runaway-optimiser-like failure modes in LLMs, where models default to unbounded single-objective maximization and neglect homeostatic targets, even after periods of initial success.
- The findings suggest that current LLMs, despite appearing multi-objective and bounded, exhibit underlying biases towards single-objective and unbounded optimization in sustained tasks.
- Uncertainty-Aware Agent Framework: introduces an LLM agent for query-conditioned multi-table summarization, leveraging retrieval uncertainty, summary uncertainty, and reinforcement learning with GRPO to filter outputs and construct high-quality synthetic datasets.
- The framework refines the agent's policy during training using reward signals based on code execution, LLM-judge scores, and summary confidence, while inference involves sampling multiple trajectories and filtering based on combined uncertainty scores.
- This approach enables agents to abstain from uncertain claims, communicate confidence, and become more reliable for complex structured-data environments, improving factuality, calibration, and downstream utility in biomedical multi-omics tasks.
When Agents go Astray: Course-Correcting SWE Agents with PRMs
- SWE-PRM (Process Reward Model): introduces an inference-time Process Reward Model that intervenes during execution to detect and course-correct trajectory-level errors, with Policy Model, Problem Description, Tool Instructions, Repository, Transcript, SWE-PRM (Process Reward Model), Taxonomy of Inefficiencies, Error Detection, Evidence Generation, Recovery Action, and Guidance Generation, where the framework prevents, detects, and course-corrects trajectory-level errors in LLM-based software engineering agents.
- The framework leverages a taxonomy of common inefficiencies to deliver lightweight, interpretable feedback without modifying the underlying policy of the LLM agent, improving reliability and efficiency in complex, multi-step software engineering tasks.
- This real-time error correction mechanism provides actionable guidance to steer the agent toward efficient completion, significantly boosting resolution rates on medium and hard tasks while reducing trajectory length.
RumorSphere: A Framework for Million-scale Agent-based Dynamic Simulation of Rumor Propagation
- RumorSphere: introduces a novel dynamic and hierarchical social network simulation framework, with an Agent Layer (distinguishes core and regular agents) comprising LLM-driven Core Agents (complex decision-making) (featuring a Persona Module (defines demographic attributes), Memory Module (stores personal environmental memory) with Personal Memory (user's historical behavior), Environmental Memory (observations and insights), Retrieval (guides behavior), Update (stores observations), and Reflection (promotes high-level thinking), an Action Module (enables agent actions), and a Belief State (represents opinion certainty)) and ABM-based Regular Agents (simpler opinion updates) (defined by Opinion (continuous belief score), fupdate (defines opinion change), fselection (determines influencing agents), and fmessage (determines transmitted message)); and an Interaction Layer (manages agent partitioning communication) that employs a Dynamic Interaction Strategy (DIS) (adaptively partitions agents) (with Adaptive Grouping (AG) (identifies core agents) and Dynamic Communication (DC) (determines communication modes)) and a Hierarchical Collaborative Network (HCN) (initializes agent network topology) (using Preferential Attachment (fosters opinion leaders) and Triangle Connection (prioritizes community links)).
- The framework supports million-scale simulations by adaptively partitioning agents into LLM-driven core agents for complex reasoning and ABM-based regular agents for efficiency, dynamically adjusting interactions based on information confusion.
- RumorSphere enables counterfactual experiments to evaluate intervention strategies, revealing that early, sustained, and opinion leader-based debunking is most effective in mitigating rumor spread within tightly connected local communities.
Batch Query Processing and Optimization for Agentic Workflows
- Halo: introduces a system for batch query processing and optimization in agentic LLM workflows, comprising a Query Parser (parses queries into DAG), a Query Optimizer (generates execution plan), and a Query Processor (executes optimized plan).
- Halo unifies query optimization with LLM serving by representing workflows as structured query plan DAGs and constructing a consolidated graph for batched queries to expose shared computation.
- The system's runtime integrates adaptive batching, KV-cache sharing and migration, and compute-communication overlap, guided by a cost model, to maximize hardware efficiency and achieve significant speedups.
JUDGEAGENT: DYNAMICALLY EVALUATE LLMS WITH AGENT-AS-INTERVIEWER
- JudgeAgent: introduces a knowledge-target adaptive dynamic evaluation framework, with Target LLM (evaluated model), Core LLM Agent (generator/evaluator LLM), Benchmark Grading (initial capability assessment), Interactive Extension (dynamic question generation/testing), Evaluation Feedback (result aggregation/suggestions), Base Datasets (static question source), Context Graph (knowledge representation), Difficulty-Adaptive Module (adjusts question difficulty), Question Synthesis Module (generates new questions), Q&A History (stores interaction history), and Evaluation Scoring Module (computes performance scores), which dynamically evaluates LLMs using an interviewer-style paradigm.
- The framework conducts comprehensive evaluations through benchmark grading, interactive extension with knowledge-driven data synthesis and target-adaptive difficulty adjustment, and provides interpretable evaluation feedback.
- JudgeAgent offers novel insights into validating evaluation methods by comparing accuracy before and after receiving suggestions, demonstrating its effectiveness in identifying and mitigating LLM knowledge and capability gaps.
Diffusion-RL Based Air Traffic Conflict Detection and Resolution Method
- Diffusion-AC: introduces a novel autonomous conflict resolution framework that integrates diffusion probabilistic models into safety-critical air traffic Conflict Detection and Resolution (CD&R), generating multimodal action distributions via a value-guided reverse denoising process, and employing a Density-Progressive Safety Curriculum (DPSC) for stable learning.
- The framework's core architecture includes a UNet-style denoising backbone with residual blocks and self-attention, a state encoder, and a time embedding module, all trained in an off-policy actor-critic fashion with dual Q-critics and target networks.
- This approach overcomes the unimodal bias of traditional DRL policies, significantly enhancing decision-making flexibility and robustness in complex, high-density air traffic scenarios, leading to a 94.1% success rate and a 59% reduction in Near Mid-Air Collisions.
- ProST (Progressive Sub-task Training): introduces a novel curriculum-style learning strategy for multi-agent systems using SLMs, with Progressive Sub-task Training Strategy (curriculum-style learning), Orchestrator Agent (decomposes tasks), Executor Agent (executes code), Critic Agent (provides feedback), AppWorld Environment (simulated app interaction), and User (initiates tasks) components, where it progressively introduces new subtasks in each training epoch to improve effectiveness and efficiency.
- This strategy addresses the challenge of SLMs struggling with long-trajectory learning by enabling them to gradually expand their learning coverage of complex problem trajectories.
- Evaluations demonstrate that ProST-trained multi-agent systems achieve superior Pareto-optimal trade-offs between effectiveness and efficiency compared to standard fine-tuning methods.
Behavioral Fingerprinting of Large Language Models
- Behavioral Fingerprinting: introduces a novel, multi-faceted framework for evaluating LLMs, including a Prompting Phase (systematically prompts LLMs), Diagnostic Prompt Suite (curated behavioral probes), Response Collection (gathers raw LLM outputs), Automated Evaluation (AI-driven response assessment), and Synthesis and Visualization (generates behavioral profiles).
- The framework employs an independent LLM as an impartial judge to assess target LLM responses against detailed rubrics, producing both quantitative visualizations and qualitative behavioral reports.
- This methodology reveals critical divergences in LLM alignment-related behaviors like sycophancy and semantic robustness, suggesting interactive nature is a direct consequence of developer strategies rather than scale or reasoning power.
DEEPTRACE: AUDITING DEEP RESEARCH AI SYSTEMS FOR TRACKING RELIABILITY ACROSS CITATIONS AND EVIDENCE
- DeepTRACE: introduces a sociotechnically grounded audit framework for evaluating deep research AI systems, with User Query (input question), Answer Text (agent's generated response with citations), Sources (listed reference URLs), Statement Decomposition (breaks answer into individual statements), Source Content Scraping (extracts full text from source URLs), LLM Judge (assigns confidence scores, determines factual support, identifies pro/con statements), Human Annotators (validates LLM judge's assessments), Citation Matrix (maps statements to cited sources), Factual Support Matrix (maps statements to factually supporting sources), One-Sided Answer Metric (measures answer bias on debate questions), Overconfident Answer Metric (measures biased confidence on debate questions), Relevant Statements Metric (measures fraction of pertinent statements), Uncited Sources Metric (measures fraction of listed but unused sources), Unsupported Statements Metric (measures fraction of claims without factual backing), Source Necessity Metric (measures fraction of essential sources), Citation Accuracy Metric (measures correctness of citations), and Citation Thoroughness Metric (measures completeness of citations), which quantifies community-identified failure cases into eight measurable dimensions for end-to-end reliability assessment.
- The framework uses statement-level analysis, confidence scoring, and builds citation and factual-support matrices to audit how systems reason with and attribute evidence, employing automated extraction pipelines for popular public models and an LLM-judge validated against human raters.
- DeepTRACE's modular design and dataset allow for flexible adaptation, enabling continuous evaluation of generative search engines (GSEs) and deep research agents (DRs) across diverse contexts, moving beyond purely technical metrics to sociotechnical impact.
IMPLICIT ACTOR CRITIC COUPLING VIA A SUPER- VISED LEARNING FRAMEWORK FOR RLVR
- PACS (imPlicit Actor Critic coupling via a Supervised learning framework): introduces a novel RLVR framework that reformulates the RLVR problem as a supervised learning task, optimizing a score function parameterized by the Policy Model (Actor/Critic) using Cross-Entropy Loss, with Reward Proxy Computation and Group Computation.
- This framework implicitly couples actor and critic roles within a single policy model, enabling more stable and efficient training by treating outcome rewards as predictable labels.
- The approach leverages reward proxy and group computations to derive advantage-like scores, demonstrating superior performance over existing RLVR baselines on challenging mathematical reasoning tasks.
OMNIACTOR: A GENERALIST GUI and EMBODIED AGENT FOR 2D&3D WORLDS
- OmniActor introduces a high-performance generalist agent for 2D and 3D worlds, utilizing a Layer-heterogeneity MoE (separates deep-layer parameters, shares shallow-layer parameters), a unified action space, and large-scale GUI and embodied data.
- The Layer-heterogeneity MoE leverages synergy between GUI and embodied data in shallow layers while eliminating conflicts in deep layers, analogous to the cerebrum-cerebellum mechanism.
- By unifying data formats and action spaces, OmniActor significantly improves performance across diverse GUI and embodied tasks, outperforming existing generalist and state-of-the-art single-environment agents.
Can Large Language Models Master Complex Card Games?
- Reviews LLMs capability to gain general capabilities through 8 diverse card games.
- Uses high quality game data to fine tune LLM and reviews its performance against specialized game AI.
- Argues the LLMs general learning capability is their largest asset compared to specialized game AI.
Structured AI Decision-Making in Disaster Management
- Structured AI Decision-Making Framework: introduces a structured decision-making framework for autonomous AI in disaster management, featuring Enabler agents (AI models providing judgment insights), Decision Maker agents (RL algorithms or human operators), Levels (critical decision points), and Scenarios (tree-like decision structures).
- The framework organizes decision flow into distinct Levels within a Scenario, where Enabler agents process disaster-related data to provide confidence scores, guiding the Decision Maker agent (either an RL algorithm or a human operator) in making informed decisions.
- The Enabler agent utilizes a Multimodal Model Architecture, combining a Text Model (BiLSTM with pooling) and an Image Model (ResNet50) to classify image-text pairs, while the RL Decision Maker agent is trained using an A2C algorithm within a custom Gymnasium environment.
LLM-empowered Agents Simulation Framework for Scenario Generation in Service Ecosystem Governance
- LLM-empowered Agents Simulation Framework for Scenario Generation in Service Ecosystem Governance: introduces a scenario generator design method, which adaptively coordinates three LLM-empowered agents—Planner Agent (PA) (coordinates schemes), Environment Agent (EA) (generates environments), and Social Agent (SA) (models agent behaviors)—along with a Data/Knowledge Base (input), Tasks (objectives), Experiment System (executes scenarios), Scenarios (outputs), and Feedback Mechanism (adjustment) to optimize experimental schemes and generate high-quality scenarios for service ecosystem governance.
- The framework leverages LLMs for semantic deconstruction, adversarial prompt engineering, and cognitive simulations to overcome limitations of predefined rules and generate diverse, extreme scenarios.
- The system's closed-loop "generate-validate-optimize" mechanism enables adaptive governance of complex service ecosystems under uncertainty, improving scenario generation efficiency and feature coverage.
Conformal Predictive Monitoring for Multi-Modal Scenarios
- GenQPM: introduces a dynamics-aware quantitative predictive monitor, with a Generative Model (learns system dynamics, generates trajectories), Mode Predictor (partitions trajectories by mode), Conformal Inference (ensures statistical guarantees), STL Robustness Calculation (quantifies property satisfaction), and Prediction Intervals (mode-specific robustness ranges), which leverages deep generative models and conformal inference for mode-specific predictive monitoring in multi-modal stochastic systems.
- This method addresses the limitation of existing quantitative predictive monitoring approaches by providing statistically valid, mode-specific prediction intervals for Signal Temporal Logic (STL) robustness, enhancing decision-making in complex dynamic environments.
- The approach offers improved interpretability and tighter prediction intervals compared to mode-agnostic baselines, enabling preemptive and timely safety interventions in systems with uncertain future behaviors.
Multi-Agent Reinforcement Learning for Task Offloading in Wireless Edge Networks
- DCC (Decentralized Coordination via CMDPs) Framework: introduces a decentralized multi-agent reinforcement learning framework for task offloading in wireless edge networks, with agents solving local CMDPs and coordinating implicitly through a shared constraint vector updated via a three-timescale learning process.
- The framework employs lightweight communication and constraint-based coupling to achieve system-level alignment while ensuring local autonomy and scalability in shared-resource environments.
- Each agent uses a reinforcement learning algorithm and Lagrange multipliers to balance individual performance objectives with global resource usage constraints, addressing challenges like non-decomposability and non-stationarity in MARL.
DeepSeek Performs Better Than Other Large Language Models in Periodontal Cases
- LLM Evaluation Framework for Periodontal Cases: introduces a system for assessing LLMs in dental case analysis, comprising a Dental Clinical Cases Collection (source of clinical data), Three-Step Prompt Design (structures LLM input), 30% Downsampling (selects subset for testing), LLMs (models being compared), Algorithm Evaluation (automated performance metrics), and Human Evaluation (expert clinical assessment).
- The framework systematically evaluates four prominent LLMs (GPT-4o, Gemini 2.0 Flash, Copilot, and DeepSeek V3) on their ability to interpret complex longitudinal periodontal case vignettes and generate professionally appropriate open-ended responses.
- DeepSeek V3 consistently demonstrated superior performance in faithfulness and expert clinical accuracy compared to other LLMs, highlighting its potential as a robust, domain-specific clinical decision-support tool for dental education and practice.
From CVE Entries to Verifiable Exploits: An Automated Multi-Agent Framework for Reproducing CVEs
- CVE-GENIE: introduces an automated, LLM-based multi-agent framework for reproducing real-world vulnerabilities from CVE entries, with all its components, where it gathers relevant resources, reconstructs the vulnerable environment, and reproduces a verifiable exploit.
- The framework's modular design, including Processor, Builder, Exploiter, and CTF Verifier, enables end-to-end CVE reproduction by specialized LLM agents, addressing challenges like incomplete data and reasoning limits through self-critique.
- CVE-GENIE successfully reproduced approximately 51% of CVEs published in 2024-2025, generating verifiable exploits and offering a robust method for creating reproducible CVE benchmarks for security research.
ShortageSim: Simulating Drug Shortages under Information Asymmetry
- ShortageSim (Large Language Model-based multi-agent simulation framework): introduces a multi-agent simulation framework for drug shortage management, featuring an Environment module, an Agents system with FDA, Manufacturer, and Buyer agents, an Information Flow, and a Simulation Controller, where each agent employs a two-stage LLM pipeline for decision-making.
- This framework models the complex, strategic interactions between drug manufacturers, institutional buyers, and regulatory agencies under information asymmetry, leveraging LLMs to simulate bounded-rational decision-making in response to shortage alerts.
- The system enables counterfactual policy analysis of FDA communication strategies and market structures, providing a novel computational framework for designing and testing interventions in information-scarce supply chains.
An LLM-enabled semantic-centric framework to consume privacy policies
- An LLM-enabled semantic-centric framework: introduces a system that automatically converts natural-language privacy policies into formal knowledge, utilizing an NLP pipeline with LLMs to identify and classify privacy-related entities and actions, and construct a Pr² Graph.
- The framework's NLP pipeline includes components for segmenting policies, recognizing and classifying data, purpose, party, and action entities, and identifying relations between them, all powered by LLMs and grounded in the Data Privacy Vocabulary (DPV).
- The resulting Pr² Graph serves as a structured representation of privacy practices, enabling downstream tasks such as constructing formal policies in ODRL or psDToU, and is publicly released for top-100 websites along with the pipeline and datasets.
In-N-Out: A Parameter-Level API Graph Dataset for Tool Agents
- In-N-Out (A Parameter-Level API Graph Dataset for Tool Agents): introduces a novel parameter-level API graph dataset, constructed through a multi-stage pipeline including documentation refinement, candidate pair filtering, and human annotation, to capture API dependencies for LLM-based tool agents.
- The dataset significantly improves tool retrieval and multi-tool query generation performance by providing explicit API dependency information, outperforming LLMs relying solely on documentation.
- Fine-tuning LLMs on In-N-Out enables them to infer parameter-level connections from documentation, generalize to unseen APIs, and achieve performance comparable to human-labeled graphs.
Cloud-Device Collaborative Agents for Sequential Recommendation
- CDA4Rec (Cloud-Device Collaborative Agents for Sequential Recommendation): introduces a novel cloud-device collaborative framework for sequential recommendation, featuring a Cloud Agent (global planning, semantic tasks) and a Device Agent (local processing, sensitive tasks), which collaboratively plan and execute personalized recommendations.
- This framework decomposes the recommendation task into sub-tasks like User Abstract Generation (summarizes user intent, behavior), Recommendation Strategy Planning (generates personalized execution plan), Semantic User Modeling (constructs intent-aware embedding), Candidate Retrieval (generates relevant item set), Structured User Modeling (captures behavioral patterns), and Final Ranking (ranks candidate items).
- CDA4Rec addresses privacy concerns, real-time responsiveness, and computational bottlenecks by dynamically assigning tasks to either the cloud-side LLM or device-side SLM based on computational demands and privacy sensitivity, ensuring efficient and adaptive personalization.
Agentic Workflow for Education: Concepts and Applications
- AWE (Agentic Workflow for Education): introduces a four-component model comprising self-reflection (iterative refinement), tool invocation (external resource use), task planning (sequential decomposition), and multi-agent collaboration (distributed intelligence), enabling dynamic, nonlinear workflows for educational applications.
- This framework distinguishes itself from traditional LLM-based linear interactions by proposing a theoretical foundation grounded in the von Neumann Multi-Agent System (MAS) architecture, shifting from static prompt-response to dynamic, nonlinear workflows.
- AWE enables scalable, personalized, and collaborative task execution across four core application domains: integrated learning environments, personalized AI-assisted learning, simulation-based experimentation, and data-driven decision-making, validated by automated math test generation.
The Need for Verification in AI-Driven Scientific Discovery
- AI-Driven Scientific Discovery Landscape: introduces a comprehensive review of computational methods for scientific discovery, encompassing traditional and AI-assisted pipelines, data-driven, knowledge-aware, derivable models, and LLM approaches, emphasizing the critical role of rigorous verification.
- The paper highlights the "verification bottleneck" in AI-assisted discovery, where rapid hypothesis generation by LLMs and other AI models outpaces the slow, manual evaluation by domain experts, hindering scientific progress.
- It advocates for improved verification methods, including automated and integrated approaches, to ensure scientific validity, interpretability, and alignment with foundational knowledge across diverse scientific domains.
DeepResearch Arena: The First Exam of LLMs' Research Abilities via Seminar-Grounded Tasks
- DeepResearch Arena: introduces a novel benchmark for evaluating LLMs' research abilities, featuring the MAHTG (Multi-Agent Hierarchical Task Generation) system, which includes Data Generation, Inspiration Extraction, Task Design, and Evaluation components, to create and assess research tasks.
- The MAHTG system processes seminar videos into transcripts, extracts categorized inspirations via an Inspira Agent and Expert Verification Team, then designs high-quality DeepResearch Tasks using TaskWeaver and RankEval Agents.
- The benchmark evaluates LLM performance through Keypoint-Aligned Evaluation (KAE) for factual correctness and Adaptively-generated Checklist Evaluation (ACE) for open-ended reasoning, both utilizing a Judge LLM.
TopoNav: Topological Graphs as a Key Enabler for Advanced Object Navigation
- TopoNav (Topological Graphs for Advanced Object Navigation): introduces a novel framework that constructs and maintains a dynamic topological memory graph as the core of its navigation system, integrating RGB-D Images, Pose, Semantic Segmentation, Semantic Point Cloud Map Construction, Topological-Based Memory Map, Current Panorama, Prompt Manager, VLM Response, Waypoint Selection Strategy, Object Detection & Verification, and Next Waypoint, to model environmental topology as actionable spatial memory for object navigation.
- This framework leverages topological structures as spatial memory, building and updating a topological graph that captures scene connections, adjacency, and semantic meaning, enabling agents to accumulate spatial knowledge, retrieve key information, and reason effectively toward distant goals.
- TopoNav achieves state-of-the-art performance in ObjectNav by connecting temporary visual inputs with lasting spatial understanding, excelling in diverse and complex environments through efficient long-horizon planning and adaptive exploration.
Aligning Requirement for Large Language Model's Code Generation
- Specine (Specification Alignment): introduces a novel specification alignment technique for LLM code generation, which identifies misaligned input specifications, lifts LLM-perceived specifications, and aligns them to generate correct code.
- The framework employs a dual-agent system with a coder agent and a tester agent for misaligned specification identification, a lifter agent for extracting LLM-perceived specifications using a Requirement Engineering DSL, and an aligner agent that applies pre-defined alignment rules to generate an aligned specification.
- Specine significantly outperforms state-of-the-art prompt-based and agent-based code generation techniques across various LLMs and benchmarks, demonstrating its effectiveness in improving code generation performance.
TableZoomer: A Collaborative Agent Framework for Large-scale Table Question Answering
- TableZoomer: introduces a novel LLM-powered, programming-based agent framework for large-scale table question answering, with a Table Describer (generates table schema), Query Planner (parses query, classifies), Table Refiner (refines schema, zooms), Code Generator (generates executable code), and Answer Formatter (formats final response), all orchestrated by a ReAct Paradigm (orchestrates iterative reasoning) and utilizing an LLM (powers agent roles) and Python Interpreter (executes generated code).
- This framework addresses TQA limitations by replacing fully verbalized tables with structured schemas, employing a query-aware table zooming mechanism for efficient data localization, and using a Program-of-Thoughts (PoT) strategy to generate executable code for robust numerical computation.
- The framework significantly enhances performance and scalability across varying table scales by reducing computational complexity and token consumption, while maintaining usability advantages through its collaborative agent design and iterative reasoning capabilities.
Communicative Agents for Slideshow Storytelling Video Generation based on LLMs
- VGTeam (Video-Generation-Team): introduces a multi-agent system for automated slideshow storytelling video generation, leveraging User Input (initial textual prompt), Chat Tower (central agent communication hub), Agent Director (coordinates agents, reviews outputs), Agent Editor (generates video captions/script), Agent Painter (generates image prompts), Agent Composer (generates music prompts), Memory Stream (stores dialogue, instructions, context), LLM API (underlying LLM capabilities for agents), Specification (defines agent roles via prompts), Text-to-Image API (generates images from text), Text-to-Speech API (generates voiceovers from text), Text-to-Music API (generates background music), Images (visual video components), Voiceover (auditory narration component), Background Music (BGM) (auditory music component), Combine Module (integrates video elements), MoviePy (video editing, post-processing tool), and Output Video (final slideshow video).
- The system employs a Chat Tower architecture for structured agent communication and an iterative approval process, ensuring quality and thematic consistency in generated video content.
- VGTeam achieves high efficiency and scalability by relying on API-driven multimedia generation and LLM role specialization, significantly reducing computational overhead and production costs.
- AnchorRAG (Multi-Agent Collaboration Framework): introduces a novel multi-agent collaboration framework for open-world Retrieval-Augmented Generation on Knowledge Graphs, featuring a predictor agent, multiple retriever agents, and a supervisor agent, all leveraging LLMs.
- The predictor agent dynamically identifies candidate anchor entities, while independent retriever agents conduct parallel multi-hop explorations on the Knowledge Graph, and the supervisor agent synthesizes knowledge paths for final answer generation.
- This framework enhances retrieval robustness and mitigates the impact of ambiguous or erroneous anchors by enabling effective knowledge retrieval without predefined anchor entities, outperforming existing baselines in real-world question answering tasks.
Web Fraud Attacks Against LLM-Driven Multi-Agent Systems
- Web Fraud Attacks: introduces a novel attack framework against LLM-driven Multi-Agent Systems (MAS) that includes IP Obfuscation, Domain Name Manipulation, Typos (Insertion, Substitution, Repetition), Subdomain Name Manipulation, Homograph Attack, Parameter Manipulation, Subdomain Imitation, Directory Imitation, and Directory Manipulation, all designed to induce MAS to visit malicious websites by exploiting link validation vulnerabilities.
- These attacks leverage structural and semantic attributes of web links to disguise malicious content as benign, requiring minimal attacker capabilities and operating from a single, low-privilege agent.
- Extensive experiments demonstrate that these attacks achieve high success rates across various MAS platforms, models, and defense strategies, highlighting a critical and overlooked vulnerability in current MAS security.
Question-to-Knowledge: Multi-Agent Generation of Inspectable Facts for Product Mapping
- Q2K (Question-to-Knowledge): introduces a multi-agent framework leveraging LLMs for reliable SKU mapping by generating and validating inspectable facts.
- This framework decomposes SKU mapping into three coordinated agents: a Reasoning Agent for targeted disambiguation questions, a Knowledge Agent for web-based evidence retrieval, and a Deduplication Agent for reusing validated reasoning traces from a Q-A Trace DB.
- Q2K incorporates a human-in-the-loop mechanism to refine uncertain cases, enhancing accuracy and robustness while reducing computational costs through efficient trace reuse.
REFRAG: Rethinking RAG based Decoding
- REFRAG (REpresentation For RAG): introduces an efficient decoding framework for RAG applications, with a Decoder-only Foundation Model (generates answers), a Light-weight Encoder (compresses context chunks), a Query Encoder (encodes user query), a Vector DB (stores retrieved embeddings), a Decoder Tokenizer & Embedding (tokenizes query input), Chunk Embedding (compressed context representation), a Light-weight RL-trained chunk expansion policy (selects chunks for expansion), and a Projection layer (matches embedding size), which compresses, senses, and expands context representations to reduce memory usage and inference latency.
- The framework leverages pre-computed, compressed chunk embeddings as approximate representations, feeding them directly into the decoder, and uses an RL policy to selectively expand crucial chunks back to full token representation.
- This approach significantly reduces time-to-first-token (TTFT) latency and memory usage by exploiting attention sparsity in RAG contexts, without requiring modifications to the underlying LLM architecture.
VERLTOOL: TOWARDS HOLISTIC AGENTIC REINFORCEMENT LEARNING WITH TOOL USE
- VERLTOOL: introduces a unified and modular framework for Agentic Reinforcement Learning with Tool Use (ARLT), featuring a Verl Workflow, a Unified API Request & Tool match, a Tool Server, and a Tool Thread, designed to disaggregate RL workflow and tool execution for efficiency and extensibility.
- The framework enables LLM Actors to engage in multi-turn rollouts, interacting with diverse tools managed by the Tool Server, which supports asynchronous execution for improved throughput and system utilization.
- Its modular plugin architecture allows rapid integration of new tools with lightweight Python definitions, providing a scalable foundation for tool-augmented RL research across various domains.
FlashAdventure: A Benchmark for GUI Agents Solving Full Story Arcs in Diverse Adventure Games
- COAST (Clue-Oriented Agent for Sequential Tasks): introduces an agentic framework for GUI agents, featuring a Clue Seeker (explores environment for clues), Clue Mapper (analyzes memory, generates subtasks), Problem Solver (executes proposed subtasks), Clue Memory (stores collected clues), Trajectory (interaction history record), and Resolved-Goal Set (completed task tracker), designed to manage long-term clue memory and solve sequential tasks in adventure games.
- The paper also introduces FlashAdventure, a benchmark of 34 Flash-based adventure games for evaluating GUI agents on full story arc completion, and CUA-as-a-Judge, an automated gameplay evaluator for reliable milestone verification.
- Experiments demonstrate that current GUI agents struggle with full story arcs due to weak planning and perception, while COAST improves milestone completion by bridging the observation-behavior gap, though a significant human-agent performance gap remains.
Plantbot: Integrating Plant and Robot through LLM Modular Agent Networks
- Plantbot: introduces a hybrid lifeform that integrates a living plant and a mobile robot through an LLM modular agent network, which includes a Living Plant (biological system), Soil (sensor-embedded substrate), Mobile Robotic Base (physical movement platform), various sensors (Soil Sensor, USB Camera, Microphone, LiDAR Sensor, Key Switch), actuators (Tracked Mobile Base, Speaker), and an LLM Modules Network comprising a Vision Agent (analyzes camera frames, suggests actions), a Sensor Agent (converts soil data to language), a Chat Agent (integrates messages, generates commands), Action Agent 1 (decides movement necessity), and Action Agent 2 (generates motor commands).
- This architecture leverages LLMs as a universal natural language protocol, translating multimodal data from biological and environmental sensors into linguistic messages that coordinate system behaviors and enable autonomous, adaptive responses.
- The system's design facilitates seamless interaction across biological and artificial domains, transforming plant states into robotic actions and installing normativity for agency within the sensor-motor loop.
Towards High Data Efficiency in Reinforcement Learning with Verifiable Reward
- DEPO (Data-Efficient Policy Optimization): introduces a two-stage data selection pipeline, with offline data selection (initial data curation) and online data selection (dynamic rollout pruning), where it combines optimized strategies for both offline and online data selection to improve data efficiency in RLVR training.
- The offline phase curates a high-quality training subset based on diversity, influence, and appropriate difficulty, while the online phase dynamically filters samples with low exploration potential and replays under-explored samples to reduce computational costs.
- This approach significantly reduces data volume and computational costs, achieving up to 1.85x speed-up on AIME24 and 1.66x speed-up on AIME25 compared to GRPO trained on the full dataset, while maintaining comparable performance.
MOBILE-AGENT-V3: FUNDAMENTAL AGENTS FOR GUI AUTOMATION
- Mobile-Agent-v3: introduces a general-purpose GUI agent framework that leverages a foundational GUI agent model, GUI-Owl, to achieve state-of-the-art performance in GUI automation across diverse desktop and mobile environments.
- The framework incorporates a self-evolving GUI trajectory production pipeline and a scalable reinforcement learning approach to continuously refine its capabilities and reduce manual annotation.
- It coordinates specialized agents for knowledge retrieval, task planning, action execution, self-correction, and persistent memory, enabling robust and adaptive long-horizon GUI task automation.
The Prompt Engineering Report Distilled: Quick Start Guide for Life Sciences
- The Prompt Engineering Report Distilled: Quick Start Guide for Life Sciences: introduces a distillation of 58 prompt engineering techniques into six core methods—zero-shot, few-shot, thought generation, ensembling, self-criticism, and decomposition—to enhance LLM performance in life sciences workflows.
- The guide provides actionable recommendations for structuring prompts, addressing common pitfalls like hallucinations and multi-turn conversation degradation, and leveraging agentic tools for improved reliability and efficiency.
- It emphasizes that systematically engineered prompts augment existing research pipelines by offering precise guidance on core principles, facilitating a transition from opportunistic to effective, low-friction systematic prompting practices.
OmniReason: A Temporal-Guided Vision-Language-Action Framework for Autonomous Driving
- OmniReason: introduces a Temporal-Guided Vision-Language-Action Framework for Autonomous Driving, comprising OmniReason-Data (VLA datasets) and OmniReason-Agent (E2E VLA model), which establishes robust spatiotemporal reasoning by jointly modeling dynamic 3D environments and their underlying decision-making processes.
- The framework addresses the limitation of existing VLMs focusing on static scene understanding by integrating explicit temporal modeling mechanisms and a hallucination-mitigated auto-labeling pipeline for data generation.
- OmniReason-Agent's architecture leverages a sparse temporal memory module and a knowledge distillation framework to internalize human-like priors and causal reasoning, enabling context-aware, interpretable, and reliable autonomous driving behavior.
- The framework simulates multi-turn conversations across Single-Visit (short-term persuasive interaction), Multi-Visit (longitudinal counseling simulation), and Social Resistance (adversarial social influence test) scenarios, with the Nurse Agent employing Direct Prompting (nurse agent response generation) or Chain-of-Strategy (CoS) (explicit strategy identification, justification) and Reflection Mechanisms (nurse agent self-critique, adaptation).
- ChatCLIDS also includes a Social Resistance Agent (simulates adversarial social influence) and uses both LLM-based Judges (automated dialogue evaluation) and Human Expert Evaluation (clinical validation of agents) for robust, multi-dimensional assessment of behavior change interventions.
Causal MAS: A Survey of Large Language Model Architectures for Discovery and Effect Estimation
- Causal MAS (Causal Multi-Agent Systems): introduces a survey of LLM architectures for causal discovery and effect estimation, featuring LLM-based agents, orchestrators/coordinators, specialized agents, debate/critique mechanisms, causal model/graph modules, interaction modules, perception modules, controllers/planners, knowledge bases/memory, statistical causal inference tools, simulation environments, and user interfaces.
- The survey explores diverse architectural patterns and interaction protocols, including pipeline-based processing, debate frameworks, simulation environments, and iterative refinement loops, to address LLM limitations in causal reasoning.
- These systems aim to enhance causal reasoning, discovery, and estimation across various application domains like scientific discovery, healthcare, and fact-checking, while tackling challenges such as hallucination and scalability.
- Aira (AI-powered Parallelization Adviser): introduces an AI-powered framework for semi-automatic fine-grained parallelization on SMT processors, including an AI Coding Agent (LLM-powered parallelization core), Cursor IDE (integrated development environment), Claude Sonnet 4 model (LLM for code analysis), Model Context Protocol (tool-LLM communication interface), sample-based profile collection (hotspot detection), Dynamic Binary Instrumentation (DBI) tool (dynamic dependency collection), binary analysis tool (static/dynamic dependency analysis), Sniper simulator (performance gain estimation), Relic parallel framework (fine-grained task execution), and a specification file (LLM workflow guidance).
- The framework integrates directly into Cursor IDE, leveraging an LLM to detect hotspots, collect dynamic dependencies, analyze static dependencies, and estimate performance gains before restructuring code with the Relic framework.
- Aira achieves an average 17% geomean performance gain for latency-critical benchmarks by enabling efficient fine-grained task parallelism on SMT cores without relying on specialized LLMs.
Supporting Our AI Overlords: Redesigning Data Systems to be Agent-First
- Agent-First Data System Architecture: introduces a new architecture for data systems designed to support LLM agent workloads, featuring an LLM Agent In Charge, MSP Agents, Field Agent, Probe Parser and Interpreter Agent, Probe answers (approx.) & grounding feedback, Sleeper Agents, Satisficing Probe Optimizer, Shared Txn Manager, Data & Metadata, and Agentic Memory Store.
- This architecture addresses the challenges of agentic speculation—high-throughput, exploratory querying by LLM agents—by leveraging its characteristics of scale, heterogeneity, redundancy, and steerability.
- The system aims to efficiently process agent "probes" (beyond SQL queries) by providing approximate answers, proactive grounding feedback, and managing shared state and memory for improved performance.
- Hybrid AI Framework: introduces a novel multi-stage hybrid intelligence framework for pruning patent portfolios, combining a Learning-to-Rank (LTR) model with a unique Need-Seed agent-based system to identify high-value assets for technology transfer.
- The framework automates and deepens patent valuation by integrating quantitative ranking based on over 30 legal and commercial parameters with qualitative market-need analysis using NLP and fine-tuned LLMs.
- It generates a "Core Ontology Framework" that matches high-potential patents (Seeds) to documented market demands (Needs), supported by a dynamic parameter weighting system and Human-in-the-Loop validation for adaptability and real-world credibility.
EVINOTE-RAG: ENHANCING RAG MODELS VIA ANSWER-SUPPORTIVE EVIDENCE NOTES
- EviNote-RAG introduces an agentic RAG framework with a structured retrieve-note-answer pipeline, including a Note-Taking Mechanism, Supportive-Evidence Notes (SENs), an Entailment Judge, Evidence Quality Reward (EQR), Reward Strategy, Policy Optimization, and Answer Generation, to enhance content distillation and reasoning reliability.
- The framework trains LLMs to compose SENs, which are concise, human-like notes preserving answer-relevant information and highlighting uncertainty, further reinforced by EQR, an entailment-based signal evaluating SENs' logical support for the final answer.
- This approach mitigates low signal-to-noise ratio and error accumulation in multi-hop reasoning, leading to improved accuracy, generalization, and training stability across various QA benchmarks.
NEWSAGENT: Benchmarking Multimodal Agents as Journalists with Real-World Newswriting Tasks
- NEWSAGENT: introduces a benchmark and agent framework for evaluating multimodal agents as journalists, enabling agents to iteratively search, edit, and rephrase content to produce news articles from real-world data.
- The framework models human journalistic workflows by providing a time-aware search function for historical context and an editing function for content modification, reflecting how human journalists gather and refine stories.
- NEWSAGENT includes 6,237 human-verified examples from real-world news events, converting multimodal content to text for broad model compatibility and evaluating LLMs on search, edit, and end-to-end newswriting capabilities.
NetGent: Agent-Based Automation of Network Application Workflows
- NetGent (Agent-Based Automation of Network Application Workflows): introduces an AI-agent framework for automating complex application workflows to generate realistic network traffic datasets, which separates workflow definition from execution by compiling natural-language rules into executable code for robust, repeatable, and efficient automation.
- The framework leverages a compile-then-replay design, utilizing a State Synthesis LLM component to generate concrete states from abstract prompts and a State Executor to deterministically replay cached code, ensuring efficiency and repeatability.
- NetGent's Web Agent integrates browser stealth, human-like interaction, and network control to achieve realism and robustness against UI variability and bot detection, enabling scalable data generation across diverse applications.
- TimeCopilot: introduces an open-source agentic framework that unifies multiple Time Series Foundation Models (TSFMs) with LLMs through a single API to automate the forecasting pipeline and provide natural language explanations.
- The framework is LLM-agnostic, supporting both commercial and open-source models, and integrates diverse forecasting families, including statistical, machine learning, and neural network methods, along with ensemble techniques.
- It streamlines the entire forecasting workflow from feature analysis and model selection to forecast generation and results explanation, enhancing reproducibility, interpretability, and accessibility.
- S³AP (Structured Social Simulation Analysis Protocol): introduces a novel formalism for representing social worlds, converting free-form narratives into structured tuples of state, observation, agent actions, and mental states, which are then used to induce Social World Models.
- The framework includes an LLM-powered S³AP Parser that transforms diverse narratives into these structured representations, enabling LLMs to better understand social dynamics and achieve state-of-the-art performance on social reasoning tasks.
- By integrating S³AP-powered Social World Models, LLM-powered AI agents can predict future social dynamics and improve decision-making, leading to more socially-aware systems capable of navigating complex social interactions.
MobiAgent: A Systematic Framework for Customizable Mobile Agents
- MobiAgent: introduces a comprehensive mobile agent system, with MobiMind-series agent models (Core agent models), AgentRR (Agent acceleration framework), MobiFlow (Benchmarking framework), and Data Collection Pipeline (Training data generation), designed to achieve state-of-the-art performance in real-world mobile scenarios.
- The MobiMind-series models employ a multi-role architecture including Planner, Decider, and Grounder for task planning, reasoning, and execution, while AgentRR accelerates performance by leveraging multi-level experiences and an ActTree structure for efficient action replay.
- MobiFlow provides a DAG-based benchmarking framework with multi-level verification mechanisms to accurately evaluate agent performance in complex mobile environments, and an AI-assisted data collection pipeline reduces manual annotation costs for training.
LLM-ASSISTED ITERATIVE EVOLUTION WITH SWARM INTELLIGENCE TOWARD SUPERBRAIN
- SuperBrain: introduces a novel framework for collective intelligence, grounded in the co-evolution of LLMs and human users, which integrates individual user-LLM dyads (Subclass Brains) with a Swarm Intelligence Layer and a Superclass Brain through bidirectional iterative evolution.
- The framework emphasizes a dynamic pathway from individual Subclass Brains, formed by persistent user-LLM interaction, to a Superclass Brain through GA-assisted forward-backward evolution and Swarm Intelligence coordination.
- This architecture provides a conceptual foundation and an architectural roadmap toward scalable, explainable, and ethically aligned collective AI, moving beyond static prompt engineering to dynamic human-LLM co-evolution.
- RESEARCHQA: introduces a resource for evaluating LLM systems by distilling survey articles from 75 research fields into 21K queries and 160K rubric items, with all components including a multi-stage pipeline, an LLM (gpt-4.1-mini) for data generation and filtering, various rubric types (survey, parametric, hybrid), expert annotators for validation, and an Ensemble Judge for evaluation.
- The framework's multi-stage pipeline systematically extracts top venues, retrieves survey articles, and generates queries and rubrics, leveraging the LLM for tasks like article classification, query refinement, and rubric item creation, ensuring data quality through extensive filtering.
- RESEARCHQA evaluates 18 parametric, retrieval-augmented, and agentic LLM systems using an Ensemble Judge that combines direct LLM preferences with rubric coverage, demonstrating significant skill gaps across systems and highlighting areas for improvement in scholarly question answering.
Exploring Decision-Making Capabilities of LLM Agents: An Experimental Study on Jump-Jump Game
- LLM Agent: introduces an architecture for an LLM-based agent to play the Jump-Jump game, comprising Perception, Reasoning, Action, and Feedback Modules, which process game state, make decisions, execute actions, and adapt strategies for optimal performance.
- The agent leverages LLMs (e.g., Claude/GPT-4) within its Reasoning Module to analyze game physics, spatial reasoning, and strategic planning, determining optimal jumping force.
- The system's performance is enhanced through systematic prompt optimization strategies, including step-by-step reasoning, few-shot learning, calibration, and error prevention, to improve decision accuracy and consistency.
- Rule-based Role Prompting (RRP): introduces a method for enhancing LLM role-playing agents, featuring improved instructions, explicit rules, persona information, and task input, further detailed by Character-Card/Scene Contract (CSC) Prompt for dialogue structuring and Hard-Enforced Function Calling (HEF) Prompt for strict tool use, to improve tool-augmented dialogue performance.
- The framework addresses common issues like over-speaking and ineffective tool use by integrating character-card/scene-contract design for structured dialogue and hard-enforced function calling for precise tool invocation.
- RRP significantly improves the effectiveness and reliability of role-playing dialogue agents, outperforming other prompting strategies in the Commonsense Persona-grounded Dialogue Challenge 2025.
Multi-Agent Data Visualization and Narrative Generation
- Multi-Agent Data Visualization and Narrative Generation System: introduces a lightweight multi-agent system that automates the data analysis workflow, from data exploration to generating coherent visual narratives for insight communication, with Data Analysis Agent (analyzes data, creates metadata), Story Generation Agent (creates narrative ideas), Story Execution Agent (ranks narratives, integrates visualizations), Visualization Generation Agent (proposes visualizations), Code Generation Agent (transforms ideas to code), Visualization Execution Agent (executes code, renders charts), Visualization Critique Agent (evaluates charts, handles errors), Report Generation Agent (selects, orders content), Report Execution Agent (renders final presentation), and Monitoring Agent (tracks system performance).
- The system combines a hybrid multi-agent architecture with deterministic components, strategically externalizing critical logic from LLMs to improve transparency and reliability, and delivering granular, modular outputs for human-AI collaboration.
- This approach uses a custom Python-based node architecture with multiprocessing to orchestrate workflows, enabling automated visual report generation with data-driven narratives from tabular datasets with minimal third-party technical dependencies.
OPEN DATA SYNTHESIS FOR DEEP RESEARCH
- InfoSeeker: introduces a scalable framework for synthesizing complex Deep Research tasks, where a Planner Agent orchestrates multi-step reasoning, a Search Engine retrieves information from a Knowledge Base, and a Refiner Agent summarizes results, all trained on the InfoSeek Dataset via Supervised Fine-Tuning and Reinforcement Learning.
- The framework addresses the scarcity of high-quality, large-scale datasets for Deep Research by generating Hierarchical Constraint Satisfaction Problems (HCSPs) with controllable complexity and verifiable answers.
- InfoSeeker-3B, a compact LLM trained with this approach, significantly outperforms larger models and commercial APIs on challenging Deep Research benchmarks.
- KG-RAG (Knowledge Graph-driven Retrieval-Augmented Generation): introduces a framework that transforms fragmented UI Transition Graphs (UTGs) into structured vector databases for efficient real-time retrieval, including UTG Extraction (xTester), Intent Generation Module (VLM, LLM), LLM Search Module (BFS, LLM Trajectory Scoring Module, Summarizer Module), and KG-RAG Knowledge Database (Structured RAG Vector Database, Retriever).
- This framework leverages an LLM-powered offline graph-search algorithm to preprocess low-quality UTGs into vector-based knowledge repositories, optimized for retrieval-augmented generation.
- During online execution, KG-RAG dynamically queries this repository using embedding-based similarity search to retrieve relevant navigational paths and app-specific information, significantly enhancing GUI agent decision-making.
LLM-Driven Policy Diffusion: Enhancing Generalization in Offline Reinforcement Learning
- LLMDPD (LLM-Driven Policy Diffusion): introduces a novel approach enhancing generalization in offline RL, with text prompts (textual task descriptions), trajectory prompts (single collected trajectories), a pre-trained LLM (processes text prompts), an MLP project head (refines text embedding), a parametric transformer (encodes trajectory prompts), a context-aware conditional policy diffusion module (policy function), a noise prediction network (estimates diffusion noise), and Q-functions (estimate cumulative reward).
- The framework leverages LLMs for rich task-relevant context from text prompts and a transformer for structured behavioral patterns from trajectory prompts, both serving as conditional inputs to the policy diffusion model.
- This integration of policy diffusion with Q-learning forms an actor-critic diffusion algorithm, enabling the RL agent to learn a generalizable, reward-maximizing policy for unseen tasks without fine-tuning.
HOW TO MAKE MUSEUMS MORE INTERACTIVE? CASE STUDY OF Artistic Chatbot
- Artistic Chatbot: introduces a voice-to-voice RAG-powered chatbot system designed to enhance visitor engagement and informal learning in cultural heritage sites, utilizing a data preprocessing pipeline and an inference pipeline for user interactions.
- The system processes raw documents through cleaning, translation, chunking, and embedding into a FAISS vector store, then uses speech-to-text, query embedding, a two-step retrieval (FAISS + CrossEncoder), an LLM for response generation, and text-to-speech for audio output.
- This chatbot adopts an artificial art curator persona, responding to free-form spoken questions in Polish, maintaining responses grounded in exhibition content, and demonstrating potential for increasing interactivity in public cultural sites.
- Collaborative Multi-Agent System (MAS): introduces an architecture for automated clinical problem detection from SOAP notes, with a Manager (orchestrates diagnostic process) coordinating dynamically assigned Specialist Agents (analyze notes, debate) powered by LLMs, using SOAP Notes (clinical input data).
- The system mimics a clinical consultation team, where the Manager dynamically assigns specialists, facilitates iterative debates among them to reach consensus, and aggregates results for a final diagnostic decision.
- This collaborative LLM architecture aims to improve diagnostic accuracy, robustness, and interpretability by surfacing and weighing conflicting evidence, outperforming single-LLM baselines in identifying clinical problems.
- YSocial (Large-Language-Model Agent Social Simulation): introduces a framework for generative social simulations, comprising a stateful platform server, a client-side simulation orchestrator, and stateless LLM services, which together enable LLM agents with persona profiles to interact within a Voat-like technology forum using a fixed catalog of technology links.
- The framework simulates a 30-day period, where LLM agents, powered by a base uncensored model (Dolphin 3.0), generate posts, replies, and reactions under platform rules, calibrated to real Voat data for operational validity.
- This approach allows for the examination of toxicity dynamics and the testing of moderation strategies in a controlled environment, demonstrating that norm-guided LLM agents can reproduce familiar online social patterns.
Cybersecurity AI: Hacking the AI Hackers via Prompt Injection
- Four-layer Defense Architecture: introduces a multi-layered defense strategy to mitigate prompt injection attacks against AI security agents, with Sandboxing & Virtualization, Primary Tool-Level Protection, File Write Protection, and Multi-Layer Validation components, aiming for complete mitigation with minimal performance overhead.
- This architecture addresses the fundamental architectural flaw in LLMs where all text in the context window is processed identically, preventing malicious instructions disguised as data from hijacking agent execution.
- The defense framework achieves 100% mitigation against various prompt injection attack vectors, demonstrating the technical feasibility of effective countermeasures despite the inherent fragility of LLM-based systems.
- LLM-Driven Optimization Architecture: introduces an integrated framework combining network optimization models with LLMs to deliver interactive, explainable, and role-aware decision support for supply chain planning, featuring a Client User Interface, REST API, AI Agents (Parser, Config Manipulator, Optimizer), LLM Models (1 & 2) for Context Engineering, Model Context Protocol, Network Optimization Model (SCIP), Bayesian Neural Network, Database, Summaries/Tables/Graphs, and a FastAPI Server.
- The system bridges the gap between complex operations research outputs and business stakeholder understanding by generating natural language summaries, contextual visualizations, and tailored key performance indicators.
- This hybrid architecture enhances decision-making confidence by translating complex optimization outcomes into clear, interactive explanations, supporting real-time interaction, configuration updates, and simulation-based insights.
- RLAIF (Reinforcement Learning from AI Feedback): introduces two AI-driven reward strategies, the Multi-Agent Rejection Sampling Framework and Adversarial Reward Signal Optimization with Reflection, to enhance Small Language Model creative writing capabilities for Chinese greetings.
- The Multi-Agent Framework generates high-quality preference data for training a reward model, while the Adversarial Framework uses a principle-guided LLM-as-a-Judge with adversarial training and reflection for direct reward signals.
- Both strategies significantly improve creative output over baselines, with the LLM-as-a-Judge approach yielding superior generation quality, training efficiency, and reduced dependency on human annotations.
- Context Management Strategies: introduces a comparison of context management strategies for LLM-based agents, including Observation Masking (replaces old observations with a placeholder) and LLM Summarization (condenses older turns into a running summary), within the SWE-agent framework.
- The study finds that a simple observation-masking strategy significantly reduces computational costs while matching or exceeding the solve rate of more complex LLM-based summarization, challenging the assumption that sophisticated context compression is always superior.
- The research highlights a "trajectory elongation" effect where LLM-based summarization can inadvertently encourage agents to persist in unproductive loops, diminishing efficiency gains despite bounded context.
- EconAgentic: introduces a Large Language Model-powered framework for analyzing Decentralized Physical Infrastructure (DePIN) markets, comprising Dynamic Market Evolution Modeling, Stakeholder Modeling and Interaction Framework, Macroeconomic Metrics for Human Value Alignment, LLM-based agents, and Heuristic-based agents.
- The framework simulates how AI agents respond to token incentives, invest in infrastructure, and adapt to market conditions, providing insights into DePIN market efficiency, inclusion, and stability.
- EconAgentic bridges the gap between industry practices and scientific research by enabling rigorous analysis and design of DePIN systems that prioritize alignment with human values at both micro and macro levels.
Think in Games: Learning to Reason in Games via Reinforcement Learning with Large Language Models
- TiG (Think-In Games): introduces a novel framework that empowers LLMs to develop procedural understanding through direct interaction with game environments, while retaining their inherent reasoning and explanatory abilities, with all components including Policy Model (LLM), Game State Representation, Macro-level Action Space, Relabeling Algorithm, GRPO (Group Relative Policy Optimization), Reference Model, Action Verifier, Reward Function, and Group Computation, where it reformulates RL-based decision-making as a language modeling task for LLMs to generate and refine language-guided policies.
- The framework leverages online reinforcement learning with environmental feedback to iteratively refine LLM-generated policies, bridging the gap between declarative and procedural knowledge in complex interactive tasks.
- TiG provides step-by-step natural language explanations for its decisions, significantly improving transparency and interpretability compared to conventional RL methods.
LLM-driven Provenance Forensics for Threat Investigation and Detection
- ProvSEEK: introduces an LLM-powered agentic framework for automated provenance-driven forensic analysis and threat intelligence extraction, designed to provide comprehensive, verifiable, and interpretable forensic investigations, which includes an LLM (orchestrates investigations), Threat Intelligence Extraction Module (converts unstructured CTI), Report Parsing Module (processes threat reports), Vector Database (stores CTI embeddings), System Database (stores logs/provenance data), Investigation Planning Agent (decomposes analysis goals), Data Retrieval Engine (executes provenance queries), Investigation Agent (aggregates artifacts, correlates), Follow-up Agent (generates follow-up steps), Safety Agent (validates actions, enforces guardrails), Explanation & Summary Module (generates human-interpretable narratives), Evidence Correlation Tools (correlates provenance data), Planning & Orchestration Tools (manages investigation workflow), and Safety & Governance Tools (validates queries, ensures safety).
- ProvSEEK leverages Retrieval-Augmented Generation (RAG) and chain-of-thought (CoT) reasoning to mitigate hallucinations and generate grounded, verifiable provenance data for forensic analysis.
- The framework achieves superior precision and recall in threat detection and intelligence extraction compared to baseline agentic AI approaches and State-Of-The-Art (SOTA) Provenance-based Intrusion Detection Systems (PIDS).
ORCA: ORchestrating Causal Agent
- ORCA (ORchestrating Causal Agent): introduces an LLM agentic system that automates end-to-end data analysis workflows in RDBMS, including an Agent Router, Data Wrangler (with Table Explorer, Table Recommender, and Text2SQL Generator), and Causal Analyzer (with Data Preparation, Config Selector, Model Implementer, and Interpreter), enabling robust data-driven decision-making with human-AI interaction.
- The framework leverages LLM-based agents to interpret user intent, retrieve and process data from external Database and Caching systems, apply causal inference techniques, and present interpretable results.
- ORCA balances automation with expert oversight through iterative human-agent interaction, allowing non-expert users to perform advanced analytical tasks without deep technical expertise.
CARJAN: Agent-Based Generation and Simulation of Traffic Scenarios with AJAN
- CARJAN: introduces a novel tool for semi-automated generation and simulation of urban traffic scenarios, integrating the AJAN multi-agent framework, the CARLA driving simulator, and a visual user interface for modeling and live simulation.
- The framework leverages SPARQL Behavior Trees for declarative, event-driven decision-making and interactions of intelligent agents, with scenarios visually modeled via a grid-based GUI and stored in an RDF triple store.
- Its carjanService middleware, built on Flask, seamlessly translates modeled scenarios into CARLA-compatible formats and executes AJAN agent commands, enabling integrated scenario testing and real-time behavior monitoring.
ReLATE: Learning Efficient Sparse Encoding for High-Performance Tensor Decomposition
- ReLATE (Reinforcement-Learned Adaptive Tensor Encoding): introduces a novel learning-augmented framework for constructing efficient sparse tensor representations, featuring a ReLATE Agent (orchestrates learning process) with an Adaptive Policy Net (learns optimal encoding policy), Adaptive Target Net (stabilizes value function estimation), Action Masking (prunes invalid actions), Action Filtering (prunes low-value actions), Reward Shaping (distributes credit for rewards), Reward Cache (stores evaluated encodings), Reward Model (predicts imagined action rewards), and Experience (stores observed environment transitions), interacting with a TD Environment (executes tensor operations) that includes Environment Representation (reduces state-action space), Encoding (sparse tensor representation), and Runtime (measures execution time).
- The framework employs an autonomous agent leveraging deep reinforcement learning and domain knowledge to discover optimized tensor encodings through direct interaction with the TD environment, learning from both real and imagined actions.
- ReLATE accelerates learning via rule-driven action masking and dynamics-informed action filtering, ensuring functionally correct tensor encoding with bounded execution time and outperforming expert-designed formats.
Instruction-Level Weight Shaping: A Framework for Self-Improving AI Agents
- ILWS (Instruction-Level Weight Shaping): introduces a lightweight framework for continual self-improvement in LLMs, treating system instructions as mutable pseudo-parameters updated post-session via reflection and user feedback, and includes a frozen LLM backbone, a Reflection Engine, a Tool Manager, a Git repository, and a Human Supervisor.
- The framework employs an LLM-driven Reflection Engine to inspect conversation traces, diagnose reasoning, and propose typed deltas (ΔSt, ΔUt, ΔTt) over instructions, user preferences, and tools, which are then score-gated, version-controlled, and optionally repaired or rolled back.
- ILWS periodically synthesizes a rating-weighted dataset from aggregated session data and distills matured instruction-space gains into the LLM's parameters, converting prompt-space improvements into weight-space without downtime.
HiVA: Self-organized Hierarchical Variable Agent via Goal-driven Semantic-Topological Evolution
- HiVA (Hierarchical Variable Agent): introduces a novel framework modeling agentic workflows as self-organized graphs with the Semantic-Topological Evolution (STEV) algorithm, which optimizes hybrid semantic-topological spaces using textual gradients as discrete-domain surrogates for backpropagation.
- The iterative process comprises Multi-Armed Bandit-infused forward routing, diagnostic gradient generation from environmental feedback, and coordinated updates that co-evolve individual semantics and topology for collective optimization in unknown environments.
- HiVA's architecture includes configurable LLM modules for agents, an evolvable tool subsystem with LLM-powered ToolGenerator and ToolUpdater, a Knowledge Graph for domain representation, and a robust modular and asynchronous architecture with sandboxed tool execution and state management.
A Whole New World: Creating a Parallel-Poisoned Web Only AI-Agents Can See
- Parallel-Poisoned Web Attack: introduces a novel attack vector leveraging a Malicious Web Server, Agent Fingerprinting Module, and Cloaking Module to serve a Cloaked Malicious Webpage with Indirect Prompt Injection to an AI Agent, while presenting a Benign Webpage to the User, thereby hijacking the agent's behavior for unauthorized actions.
- This stealthy attack exploits the unique digital fingerprints of web-browsing LLM agents, making it invisible to human users and conventional security crawlers, and enabling data exfiltration, malware execution, or misinformation propagation.
- The attack turns the victim's own trusted AI Agent into an attack tool by overriding its original goals with hidden instructions, demonstrating a critical security paradigm shift for autonomous web agents.
- AI-Generated Social Simulations (Methodological Docking Experiment): introduces a comparative validation study, with Human Founders (qualitative interview data source), Synthetic Users (computational simulation actors), SyntheticUsers.com platform (generates synthetic agents), Interview Protocol (mirrors human study scope), Thematic Analysis (codes transcript data), and Comparative Framework (evaluates simulation fidelity), designed to assess the credibility of LLM-driven personas as social simulation agents for startup validation research.
- The SyntheticUsers.com platform, a core component, leverages an ensemble-style routing agent to dynamically shuffle between multiple LLMs, integrates personality frameworks and affective modeling for human-like responses, and uses a RAG layer with behavioral datasets for domain-specific and demographically aligned outputs.
- This framework systematically aligns human-subject data with synthetic agents to evaluate convergence, divergence, and blind spots, positioning LLM-driven personas as a hybrid simulation category that extends traditional agent-based models with linguistic richness and psychological nuance.
Democratizing Agentic AI with Fast Test-Time Scaling on the Edge
- FlashTTS: introduces a serving system for Test-Time Scaling (TTS) on edge devices, with Speculative Beam Extension (hides straggler latency), Dynamic Prefix-Aware Scheduling (maximizes KV-cache reuse), and Asymmetric Multi-Model Memory Allocation (balances generator/verifier memory), built on vLLM.
- This framework enables edge LLMs (≤ 7B) to achieve accuracy and latency comparable to large cloud models by addressing hardware underutilization, suboptimal KV cache reuse, and memory pressure from multi-model execution.
- FlashTTS significantly improves goodput and reduces latency by leveraging a two-phase scheduling policy, roofline-guided KV allocation, and extended search space with offloading to make agentic AI practical on memory-constrained edge devices.
HiVA: Self-organized Hierarchical Variable Agent via Goal-driven Semantic-Topological Evolution
- HiVA (Hierarchical Variable Agent): introduces a novel framework modeling agentic workflows as self-organized graphs with the Semantic-Topological Evolution (STEV) algorithm, which optimizes hybrid semantic-topological spaces using textual gradients as discrete-domain surrogates for backpropagation.
- The iterative process comprises Multi-Armed Bandit-infused forward routing, diagnostic gradient generation from environmental feedback, and coordinated updates that co-evolve individual semantics and topology for collective optimization in unknown environments.
- HiVA's architecture includes configurable LLM modules for agents, an evolvable tool subsystem with LLM-powered ToolGenerator and ToolUpdater, a Knowledge Graph for domain representation, and a robust modular and asynchronous architecture with sandboxed tool execution and state management.
CoComposer: LLM Multi-agent Collaborative Music Composition
- CoComposer (LLM Multi-agent Collaborative Music Composition): introduces a multi-agent system for collaborative music composition, featuring five specialized LLM-based agents, AutoGen for collaboration, and a MIDI backend for sound generation.
- The system addresses limitations in AI music composition by closely mimicking traditional music workflows, enhancing music quality, production complexity, and controllability.
- CoComposer, which uses ABC notation as an intermediate carrier, demonstrates improved interpretability and editability compared to non-LLM models, despite MusicLM's superior aesthetic quality.
- DR.INFO (Agentic RAG-based clinical support assistant): introduces an agentic, RAG-based clinical support assistant, with an Agentic component (enables complex reasoning), a RAG-based component (retrieves and augments responses), and an LLM (generates responses), evaluated using HealthBench, a rubric-driven benchmark composed of open-ended, expert-annotated health conversations.
- HealthBench provides a multi-dimensional evaluation framework with physician-authored rubrics, themes, and behavioral axes to assess LLM performance in realistic clinical scenarios, moving beyond traditional multiple-choice benchmarks.
- The evaluation demonstrates DR.INFO's strengths in communication, instruction following, and accuracy, outperforming frontier LLMs and other agentic RAG assistants on the HealthBench Hard subset, while also identifying areas for improvement in context awareness and completeness.
HiVA: Self-organized Hierarchical Variable Agent via Goal-driven Semantic-Topological Evolution
- HiVA (Hierarchical Variable Agent): introduces a novel framework modeling agentic workflows as self-organized graphs with the Semantic-Topological Evolution (STEV) algorithm, which optimizes hybrid semantic-topological spaces using textual gradients as discrete-domain surrogates for backpropagation.
- The iterative process comprises Multi-Armed Bandit-infused forward routing, diagnostic gradient generation from environmental feedback, and coordinated updates that co-evolve individual semantics and topology for collective optimization in unknown environments.
- HiVA's architecture includes configurable LLM modules for agents, an evolvable tool subsystem with LLM-powered ToolGenerator and ToolUpdater, a Knowledge Graph for domain representation, and a robust modular and asynchronous architecture with sandboxed tool execution and state management.
HiVA: Self-organized Hierarchical Variable Agent via Goal-driven Semantic-Topological Evolution
- HiVA (Hierarchical Variable Agent): introduces a novel framework modeling agentic workflows as self-organized graphs with the Semantic-Topological Evolution (STEV) algorithm, which optimizes hybrid semantic-topological spaces using textual gradients as discrete-domain surrogates for backpropagation.
- The iterative process comprises Multi-Armed Bandit-infused forward routing, diagnostic gradient generation from environmental feedback, and coordinated updates that co-evolve individual semantics and topology for collective optimization in unknown environments.
- HiVA's architecture includes configurable LLM modules for agents, an evolvable tool subsystem with LLM-powered ToolGenerator and ToolUpdater, a Knowledge Graph for domain representation, and a robust modular and asynchronous architecture with sandboxed tool execution and state management.
HiVA: Self-organized Hierarchical Variable Agent via Goal-driven Semantic-Topological Evolution
- HiVA (Hierarchical Variable Agent): introduces a novel framework modeling agentic workflows as self-organized graphs with the Semantic-Topological Evolution (STEV) algorithm, which optimizes hybrid semantic-topological spaces using textual gradients as discrete-domain surrogates for backpropagation.
- The framework includes agent semantics (LLMs with prompts/tools), a Semantic-Topological Evolution (STEV) algorithm (core optimization algorithm), Multi-Armed Bandit-infused Forward Routing (KABB) (dynamic agent selection/routing), Textual Gradient Feedback (language-based diagnostic signals), and an Aggregator (synthesizes outputs/generates answers).
- HiVA's iterative process comprises Multi-Armed Bandit-infused forward routing, diagnostic gradient generation from environmental feedback, and coordinated updates that co-evolve individual semantics and topology for collective optimization in unknown environments.
Ultron: Foundational GUI Agent with Advanced Perception and Planning
- Ultron (Foundational GUI Agent with Advanced Perception and Planning): introduces an open-source foundational GUI agent framework with a robust data engineering pipeline, a unified interactive infrastructure, and a three-stage training paradigm, enabling advanced GUI perception, grounding, and planning capabilities.
- The framework leverages extensive data engineering strategies, including multi-turn conversations, multi-task unification, and cross-platform generalization, to enhance training effectiveness and model understanding across diverse GUI scenarios.
- Ultron significantly improves interactive capabilities in Chinese mobile app scenarios through meticulously collected operation trajectories and tailored online environments, demonstrating substantial progress toward real-world GUI agent deployment.
- Conversation-Tree Recipe (Structured-Prompting): introduces a framework for designing smarter conversational agents for children, with components including a Large Language Model (LLM) via OpenAI API, System Boundaries, Mode Boundaries, Learning Customization, Learning Assessment, and Game Generation, to enhance scaffolded learning and engagement.
- This recipe constrains LLMs to generate grade-appropriate, pedagogically scaffolded dialogue by dynamically adjusting interaction based on a child's grade level, mode (school, discovery, entertainment), and knowledge level.
- The framework aims to blend human-human and human-computer communication principles, supporting critical thinking, problem-solving, and seamless transitions between various child activities.
BED-LLM: INTELLIGENT INFORMATION GATHERING WITH LLMS AND BAYESIAN EXPERIMENTAL DESIGN
- BED-LLM (Bayesian Experimental Design with Large Language Models): introduces a general-purpose approach for improving LLMs' ability to intelligently and adaptively gather information from a user or external source using sequential Bayesian experimental design, including LLMs (core intelligent agents), Sequential Bayesian Experimental Design (guiding iterative framework), Expected Information Gain (EIG) Maximization (question selection criterion), Probabilistic Model (represents beliefs, generative process), LLM's Belief Distribution (internal uncertainty representation), EIG Estimator (calculates information gain), Candidate Query Generation Strategy (proposes diverse questions), History (ht) (accumulated past interactions), User/External Source (provides responses), Prior-likelihood pairing (joint model construction), Rejection Sampling Procedure (filters belief samples), Hypothesis-retention mechanism (maintains consistent hypotheses), Questioner LLM (asks questions), Answerer LLM (simulates user responses), and LLM-as-judge protocol (evaluates recommendations).
- The framework integrates LLMs as core intelligent agents, employing a carefully designed EIG estimator, a targeted candidate query generation strategy, and a robust model updating mechanism including rejection sampling and hypothesis retention.
- BED-LLM significantly outperforms direct LLM prompting and other adaptive design strategies in tasks like 20-Questions and active preference elicitation, demonstrating its effectiveness in multi-turn conversational and interactive environments.
A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers
- Sci-LLMs (Scientific Large Language Models): introduces a three-stage evolutionary framework for AI in scientific research, encompassing Data Foundation (foundational data infrastructure, efficient data processing, diverse data handling, continuous knowledge integration, data quality assessment), Scientific Knowledge Emergence (scientific capabilities, broad applicability, logical problem-solving, understandable decision-making), and Agent-driven Scientific Discovery (autonomous AI agents, self-directed research execution, governance, fairness, privacy, closed-loop data feedback).
- This framework outlines the progression from foundational data infrastructure and emerging scientific capabilities to autonomous AI agents capable of self-evolving discovery systems.
- The survey emphasizes the co-evolution of models and their underlying data substrate, providing a roadmap for building trustworthy and continually evolving AI systems for scientific discovery.
- LLM-driven Price Negotiation Simulation Framework: introduces a system to investigate cognitive biases in LLMs, with Seller Agent (Large Language Model), Buyer Agent (Large Language Model), Personality Profiles, Anchoring Effect Module, Reasoning Module, Dialogue System, Objective Metric, Subjective Metric, Susceptibility Metric, Prompt Settings, and Negotiation Scenarios.
- The framework simulates price negotiations between LLM agents, assessing the anchoring effect's influence through objective utility and subjective satisfaction metrics, while also exploring the roles of reasoning and personality traits.
- Findings indicate that LLMs are susceptible to the anchoring effect similar to humans, reasoning can mitigate this bias, and no significant correlation exists between personality traits and anchoring susceptibility.
- ChatThero: introduces an LLM-supported chatbot for behavior change and therapeutic support in addiction recovery, featuring a Patient Profile (structured patient characteristics), Dynamic Memory (evolving patient state), Multi-Agent Simulation Framework (generates synthetic dialogues), Patient Agent (GPT-4o-mini) (simulates patient behavior), Therapy Agent (ChatThero) (deploys therapeutic strategies), Environment Agent (introduces external stressors), Therapeutic Strategies (clinically validated approaches), SFT Dataset (supervised fine-tuning data), DPO Dataset (preference optimization data), Supervised Fine-Tuning (initial model training), Direct Preference Optimization (refines therapeutic behaviors), Human Evaluators (clinical expert feedback), and AI Evaluators (GPT-4o) (automated feedback), designed to provide scalable, adaptive, and ethical support for addiction recovery.
- The framework utilizes a two-stage training pipeline, comprising supervised fine-tuning (SFT) followed by direct preference optimization (DPO), to refine persuasive strategies based on expert and AI feedback.
- ChatThero consistently outperforms baselines across patient difficulty levels, demonstrating greater resilience and communicative effectiveness in challenging scenarios, and is rated higher in empathy, responsiveness, and behavioral realism by human and automated clinical assessments.
ProactiveEval: A Unified Evaluation Framework for Proactive Dialogue Agents
- ProactiveEval: introduces a unified evaluation framework for proactive dialogue agents, which decomposes proactive dialogue into target planning and dialogue guidance tasks, establishing evaluation metrics across various domains and enabling automatic generation of diverse evaluation data.
- The framework leverages a hierarchical environment topic tree, target ensemble techniques, and adversarial strategies like obfuscation rewriting and noise injection to synthesize challenging evaluation data.
- It employs an LLM-as-a-judge method for comprehensive assessment and utilizes a simulated user for interactive dialogue guidance evaluation.
- IRMA (Input-Reformulation Multi-Agent): introduces a verification-loop-free framework that enhances the input for a tool-calling LLM agent by reformulating user queries with structured and contextually relevant information, including Memory Module (stores conversation history), Constraints Module (generates domain policies), and Tool Suggestion Module (generates relevant tool list).
- This framework guides the LLM agent to better adhere to domain policies and improve tool selection by enriching its input with key constraints and tool-related context, leading to improved agent behavior.
- IRMA significantly outperforms other methods like ReAct, Function Calling, and Self-Reflection in terms of accuracy, reliability, and efficiency in complex, dynamic multi-turn conversational environments.
PromptSleuth: Detecting Prompt Injection via Semantic Intent Invariance
- PromptSleuth: introduces a semantic-oriented defense framework for detecting prompt injection, with a Summarization Module (extracts abstract tasks), a Task-relationship Graph Generation Module (models semantic relationships), a Clustering Module (consolidates related tasks), a Detection Module (identifies prompt injection), and an internal Detector LLM (task summarizer, relationship analyzer).
- This framework identifies prompt injection attacks by reasoning over task-level intent and logical inconsistencies, rather than relying on surface-level cues.
- PromptSleuth generalizes by identifying invariant malicious intent despite evolving attack variants, offering a robust, efficient, and generalizable strategy for safeguarding LLMs.
cMALC-D: Contextual Multi-Agent LLM-Guided Curriculum Learning with Diversity-Based Context Blending
- cMALC-D (Contextual Multi-Agent LLM-Guided Curriculum Learning with Diversity-Based Context Blending): introduces a framework that leverages an LLM (Large Language Model) to dynamically generate semantically meaningful curricula for MARL agents, using a context buffer and a diversity-based context blending mechanism.
- The framework adaptively proposes new environment contexts by reasoning over context variables and agent learning progress, preventing mode collapse and encouraging exploration through context blending.
- Experiments in traffic signal control domains demonstrate that cMALC-D significantly improves generalization and sample efficiency compared to existing curriculum learning baselines.
- AICL (Agent Interaction Communication Language): introduces a structured protocol for testable LLM applications, with components including HELLO (session initialization, handshake), QUERY (request to agent/tool), PLAN (multi-step reasoning/execution plan), FACT/FACTS (known information, environmental conditions), RESULT (output for QUERY/PLAN), ERROR (standardized error reporting), MEMORY.STORE (explicitly stores state/information), MEMORY.RECALL (retrieves stored information), COORD.DELEGATE (delegates subtask to agent/tool), and REASONING.(START|STEP|COMPLETE) (marks structured reasoning stages).
- The paper decomposes LLM applications into a three-layer architecture (System Shell Layer, Prompt Orchestration Layer, LLM Inference Core) to analyze testing applicability and proposes four collaborative strategies (Retain, Translate, Integrate, Runtime) for a trustworthy quality assurance framework.
- AICL operationalizes these strategies by enforcing semantic precision, encoding observability and provenance, guaranteeing replayability, and providing built-in evaluation hooks for automated verification and systematic failure analysis in LLM application testing.
RE⁴: SCIENTIFIC COMPUTING AGENT WITH REWRITING, RESOLUTION, REVIEW AND REVISION
- RE⁴ (Scientific Computing Agent with Rewriting, Resolution, Review and Revision): introduces a novel agent framework for scientific computing, with Consultant LLM, Programmer LLM, and Reviewer LLM collaborating through a rewriting-resolution-review-revision logical chain.
- This multi-LLM collaborative framework significantly improves bug-free code generation and reduces non-physical solutions by iteratively refining code through interactive feedback from runtime outputs.
- The agent framework demonstrates generality and versatility by successfully solving PDEs, ill-conditioned linear systems, and data-driven physical analysis problems.
CyberSleuth: Autonomous Blue-Team LLM Agent for Web Attack Forensics
- CyberSleuth (Autonomous Blue-Team Large Language Model Agent for Web Attack Forensics): introduces an autonomous LLM agent designed for the forensic investigation of web application attacks, processing packet-level traces and application logs to identify targeted services, exploited vulnerabilities (CVEs), and attack success, and generating structured forensic reports.
- The framework employs a multi-agent architecture, specifically the Flow Reporter Agent (FRA) design, which includes a Main Agent coordinating with specialized sub-agents like the Flow Summariser and Log Summariser, and external tools such as a Web Search Tool, all supported by an LLM Backend and MemGPT-style memory management.
- CyberSleuth's design emphasizes simple orchestration over complex inter-agent communication and highlights the importance of balanced data processing, demonstrating improved CVE identification accuracy and providing a benchmark for evaluating defensive LLM agents.
GDS Agent: A Graph Algorithmic Reasoning Agent
- GDS Agent (Graph Data Science agent): introduces a system for graph algorithmic reasoning, with a User (initiates questions), LLM (MCP client, generates tool calls, final answer), MCP (Model Context Protocol) Server (core agent, hosts tools, connects database), Neo4j Database (stores graph data), GDS (Graph Data Science) Library (provides graph algorithms), Tools (graph algorithms, auxiliary functions), Cypher Projection (creates in-memory subgraph), Projected Graph (in-memory graph for algorithms), Preprocessing (retrieves relevant data), and Postprocessing (formats algorithm results).
- The agent enables LLMs to perform complex graph algorithmic reasoning on large-scale knowledge graphs by integrating a comprehensive set of GDS algorithms as tools within an MCP server, allowing for accurate and grounded answers to user questions.
- This framework addresses the limitation of LLMs in directly processing graph-structure data, amplifying their utility for analyzing private or enterprise knowledge graphs and simplifying access to graph analytics libraries.
SemSR: Semantics aware robust Session-based Recommendations
- SemSR (Semantics aware robust Session-based Recommendations): introduces a framework for session-based recommendations that integrates LLM-generated semantic embeddings with data-driven SR models, including an SR Model, LLM, Attention Layer, Linear Layer, Concatenation, Cosine Similarity, Softmax, and a Trainable Embedding Look-up Table.
- The framework offers two main variants: SemSR-F, which fuses LLM-based item and session embeddings with data-driven representations, and SemSR-I, which initializes SR models with LLM-generated item embeddings.
- SemSR aims to enhance recommendation performance by leveraging the semantic understanding capabilities of LLMs to complement traditional collaborative information from data-driven SR models, leading to improved recall and MRR metrics.
MCP-Bench: Benchmarking Tool-Using LLM Agents with Complex Real-World Tasks via MCP Servers
- MCP-Bench (Benchmarking Tool-Using Large Language Model Agents with Complex Real-World Tasks via Model Context Protocol Servers): introduces a benchmark for evaluating LLM agents on realistic, multi-step tasks, featuring Real-world MCP Servers (expose 250 structured tools), LLM-based Task Synthesis (generates complex, fuzzy tasks), an LLM Agent (executes multi-step tool invocations), Execution Results and Trajectory (records agent's actions), Rule-based Evaluation (checks tool validity, schema, runtime), LLM-as-a-Judge Evaluation (scores task completion, planning), and Agent Performance (measures overall agent capability).
- This benchmark connects LLM agents to 28 live MCP servers across diverse domains, enabling the creation of authentic multi-step tasks that require tool use, cross-tool coordination, and precise parameter control, which are then evaluated using a multi-faceted framework.
- MCP-Bench addresses limitations of prior API-based benchmarks by focusing on fuzzy instructions, multi-hop execution, information grounding, and cross-domain orchestration, revealing persistent challenges for advanced LLMs in complex tool-using scenarios.
- MINDGUARD: introduces a decision-level guardrail for LLM agents, providing provenance tracking of call decisions, policy-agnostic detection, and poisoning source attribution against Tool Poisoning Attacks (TPA).
- It operates by parsing the LLM's context, building a Decision Dependence Graph (DDG) from attention matrices, and analyzing the DDG to detect and attribute poisoned invocations.
- The framework is non-invasive, explainable, and operates in real-time without modifying the underlying LLM, achieving high accuracy in detecting poisoned invocations and attributing their source.
CAPE: Context-Aware Personality Evaluation Framework for Large Language Models
- CAPE (Context-Aware Personality Evaluation) Framework: introduces a novel evaluation approach for LLMs, with Large Language Models (LLMs), Conversational History, Psychometric Tests, Inconsistency Factors, Trajectory Consistency (TC) Metric, OCEAN Consistency (OC) Metric, Gaussian Process Regression (GPR), and Role Playing Agents (RPAs), where it evaluates LLM personality by incorporating prior conversational interactions to assess response consistency and personality shifts.
- The framework utilizes psychometric tests and introduces novel metrics, Trajectory Consistency (TC) and OCEAN Consistency (OC), to quantify LLM response consistency under various prompt sensitivity factors like temperature and option wording.
- The framework demonstrates that conversational history enhances response consistency through in-context learning but can also induce personality shifts in LLMs, particularly when applied to Role Playing Agents.
Adaptive Root Cause Localization for Microservice Systems with Multi-Agent Recursion-of-Thought
- RCLAgent (Adaptive Root Cause Localization for Microservice Systems with Multi-Agent Recursion-of-Thought): introduces an adaptive root cause localization method for microservice systems, with a Coordinator (orchestrates phases), Data Agents (retrieve/process trace, metric, and format data), and Thought Agents (perform recursive and intermodal inference reasoning).
- The framework employs a novel recursion-of-thought strategy to guide the LLM's reasoning process, effectively integrating data from multiple agents and tool-assisted analysis to accurately pinpoint the root cause.
- RCLAgent achieves superior performance by localizing the root cause using only a single request, outperforming state-of-the-art methods that depend on aggregating multiple requests.
AI-SEARCHPLANNER: MODULAR AGENTIC SEARCH VIA PARETO-OPTIMAL MULTI-OBJECTIVE REINFORCEMENT LEARNING
- AI-SearchPlanner: introduces a novel reinforcement learning framework designed to enhance end-to-end QA performance by decoupling search planning from answer generation and optimizing it via multi-objective reinforcement learning.
- The framework offloads QA functionality to a large, frozen Generator LLM, while a smaller, trainable Search Planner LLM focuses on search planning, ensuring flexibility and efficiency for real-world applications.
- It employs a dual-reward mechanism for search planning, aligning outcome-level performance gains and process-level trajectory rationality, while Pareto optimizing planning utility and computational cost.
Multi-Agent Penetration Testing AI for the Web
- MAPTA (Multi-Agent Penetration Testing AI): introduces a multi-agent system for autonomous web application security assessment, with Coordinator Agent (LLM-driven, orchestrates strategy, delegates), Sandbox Agent(s) (LLM-driven, executes tactical commands), Validation Agent (LLM-driven, verifies PoC exploits), Per-Job Docker Container (isolated execution environment), Target Web App (application under security assessment), Usage Tracker (monitors resources, enforces budgets), and PoC Storage (stores candidate exploit artifacts).
- This framework combines LLM orchestration with tool-grounded execution and end-to-end exploit validation to bridge the semantic gap between vulnerability detection and contextual exploitation.
- MAPTA transforms security assessment from human-dependent pattern recognition to adaptive adversarial execution, enabling autonomous reasoning and validation at machine scale.
rStar2-Agent: Agentic Reasoning Technical Report
- rStar2-Agent: introduces a 14B math reasoning model trained with agentic reinforcement learning, incorporating a scalable RL Infrastructure, an Environment Service, the GRPO-RoC (Group Relative Policy Optimization with Resampling on Correct) RL algorithm, a Python code environment, a Tool call interface, a Prompt Template, a Math-Verifier tool, a Non-reasoning SFT stage, and Multi-stage RL training, to achieve frontier-level performance in math reasoning.
- The framework's GRPO-RoC algorithm, with its Resample-on-Correct rollout strategy, effectively addresses environment noise from coding tools by filtering positive trajectories for minimal errors and uniformly downsampling negative ones, improving training stability and reasoning quality.
- The efficient RL infrastructure, featuring a load-balanced rollout scheduler and a high-throughput isolated code environment, enables training on limited GPU resources by maximizing computational utilization and handling massive concurrent tool calls with low latency.
HCQA: Hybrid Classical-Quantum Agent for Generating Optimal Quantum Sensor Circuits
- HCQA (Hybrid Classical-Quantum Agent): introduces a hybrid AI-quantum framework for generating optimal Quantum Sensor Circuits (QSCs), integrating a DQN for policy optimization and a quantum-based action selection mechanism, where QFI serves as the reward signal.
- The framework leverages a quantum circuit to encode the agent's state using Ry gates, create action superpositions with H gates, and measure for probabilistic action outcomes, guided by Q-values.
- This approach efficiently produces entangled quantum states by selecting sequences of Rx, Ry, and S gates that maximize QFI while minimizing gate complexity, enhancing quantum metrology and control tasks.
Adaptive Monitoring and Real-World Evaluation of Agentic AI Systems
- AMDM (Adaptive Multi-Dimensional Monitoring): introduces a practical algorithm for real-time evaluation of agentic AI systems, which processes streaming metrics through normalization and aggregation into five evaluation axes, applies adaptive EWMA thresholds for per-axis anomaly detection, and performs joint anomaly detection using Mahalanobis distance to trigger mitigation or human review.
- The framework significantly reduces anomaly detection latency and false-positive rates compared to static thresholds by dynamically adapting to metric distributions and identifying multi-dimensional deviations.
- AMDM transforms a conceptual five-axis evaluation framework into an operational tool, enabling balanced monitoring of agentic AI systems across technical, human-centered, and economic dimensions to surface issues like goal drift, safety violations, and trust shocks.
- LLM-RL Integration Pipeline: introduces a framework that integrates an LLM into the RL pipeline to augment decision-making in Autonomous Cyber Operations, featuring CybORG Raw Output, CybORG State Preprocessing, Prompt Construction, LLM (Frozen), LLM Response Generation, LLM Recommendation Extraction, LLM Recommendation Mapping, RL Agent, Actor Network, Masking Function, Experience Buffer, CybORG Environment, Reward Signal, PPO Loss Function, Auxiliary Loss Function, Total Loss, and Wrapper.
- This pipeline guides the RL agent's training by leveraging a pretrained LLM as a teacher, using action masking during inference and an auxiliary loss signal during training to incorporate external cybersecurity knowledge.
- The approach improves training efficiency, reduces the need for suboptimal exploratory actions, and accelerates convergence to a favorable policy in simulated cybersecurity environments.
Operating advanced scientific instruments with AI agents that learn on the job
- AG2 (Autogen framework): introduces a human-in-the-loop pipeline for operating advanced scientific instruments, featuring a multi-agent system powered by LLMs, including specialized agents for code generation, review, administration, information extraction, image analysis, and teachability, alongside core capabilities for planning, actions, tools, and memory management.
- This framework integrates human input and iterative learning to orchestrate complex, multi-task scientific workflows, interpret multimodal data, and interactively collaborate with human researchers.
- The system demonstrates continuous learning from human feedback, storing past interactions in a vector database to enhance adaptability and improve performance in robotic control sequences.
Divide, Discover, Deploy: Factorized Skill Learning with Symmetry and Style Priors
- D3 (Divide, Discover, Deploy): introduces a modular Unsupervised Skill Discovery (USD) framework, with Environment, Data Collection Module, Skill-Conditioned Policy, Skill Prior, Factor Weighting Prior, Skill Discovery Reward Module (including METRA Algorithm, DIAYN Algorithm, Style Reward), On-Policy RL Training Module, Symmetry Augmentation Module, Intrinsic Reward Module, Value Function Decomposition Module, Advantage Aggregation Module, Training Module, Factorized State Space, Factorized Skill Space, Factor Weights, and Regularization Penalties, which addresses safety, interpretability, and deployability challenges in learned skills by factorizing the state space and applying tailored USD algorithms with symmetry and style priors.
- The framework leverages user-defined factorization of the state space, assigning specific USD algorithms (METRA or DIAYN) to each factor, and incorporates symmetry-based inductive biases and a style factor to promote structured, morphology-aware, safe, and robust behaviors.
- D3 further enhances control and coordination through factor weighting, allowing dynamic prioritization of skill components, and demonstrates zero-shot transfer of learned quadrupedal skills from simulation to real hardware.
- AgentCoMa (Agentic Commonsense and Math benchmark) introduces a compositional benchmark for LLM agents, featuring compositional questions (tasks requiring both commonsense and mathematical reasoning), commonsense reasoning steps (initial choice based on everyday knowledge), mathematical reasoning steps (subsequent arithmetic operation), real-world agentic scenarios (five practical domains), evaluation metrics (accuracy on steps and composition), analysis components (neuron patterns, attention maps, membership inference), and benchmarked LLMs (61 diverse models).
- This benchmark reveals a significant compositionality gap in LLMs, where models achieve high accuracy on isolated commonsense and math steps but experience a substantial performance drop when these mixed-type steps are combined in compositional tasks.
- Interpretability analyses indicate that LLMs struggle with mixed-type reasoning due to the rarity of such tasks in their training data, leading to the activation of neural circuits relevant to only one reasoning type during compositional problem-solving.
- Multi-Agent Framework for Ophthalmic Surgical Planning: introduces an AI-driven system for cataract surgery planning, featuring a Knowledge-driven Multi-Agent System (MAS) for report interpretation, the CataractSurg-80K Dataset for structured reasoning, and the Qwen-CSP Model for clinical decision support.
- The MAS employs collaborative specialist agents to process Raw Ophthalmic Reports into structured Patient Descriptions, simulating expert Doctor Reasoning for transparent data extraction.
- The Qwen-CSP Model, built on a Base LLM (Qwen3-4B), undergoes Multi-Stage Domain-Aware Fine-Tuning using Clinical Knowledge and Real Medical Data from the CataractSurg-80K Dataset to optimize ophthalmic surgical reasoning.
CASE: An Agentic AI Framework for Enhancing Scam Intelligence in Digital Payments
- CASE (Conversational Agent for Scam Elucidation): introduces a novel Agentic AI framework for enhancing scam intelligence in digital payments, featuring a Conversational Agent (user-facing interaction) and an Information Extractor Agent (processes transcripts), designed to collect and manage user scam feedback in a safe and scalable manner.
- The framework's Conversational Agent proactively interviews potential victims to elicit detailed scam intelligence, which the Information Extractor Agent then processes into structured data for downstream enforcement mechanisms.
- Implemented on Google Pay India using Gemini LLMs, the framework demonstrated a 21% uplift in scam enforcement volume and significantly improved response speed to new threats.
Your AI Bosses Are Still Prejudiced: The Emergence of Stereotypes in LLM-Based Multi-Agent Systems
- Multi-Agent Simulation Framework: introduces a novel experimental framework to investigate stereotype emergence and evolution in LLM-based multi-agent systems, simulating workplace interactions with LLM-Based Agents, a Supervisor Agent, and dedicated Evaluation and Parser Agents, all interacting through defined cycles and maintaining a comprehensive interaction history.
- The framework employs synchronized task-interaction cycles, allowing for both random and hierarchical task assignments, and quantifies stereotype formation using specialized metrics across diverse LLM architectures.
- This design enables the study of how stereotypes emerge spontaneously in AI agent interactions, intensify with increased interaction rounds and decision-making power, and manifest consistently across different LLM architectures.
Secure Multi-LLM Agentic AI and Agentification for Edge General Intelligence by Zero-Trust: A Survey
- The Zero-Trust Multi-LLM Framework (ZT-MLLMF): introduces a comprehensive survey of zero-trust security principles applied to multi-LLM systems in Edge General Intelligence (EGI), detailing architectural design and operational workflows.
- The paper systematically analyzes critical security vulnerabilities in collaborative multi-LLM systems, including insecure inter-LLM communications and expanded attack surfaces, which traditional perimeter-based security cannot adequately address.
- ZT-MLLMF implements zero-trust principles such as explicit verification, least privilege, continuous monitoring, and micro-segmentation through model- and system-level approaches to enhance security and trustworthiness.
Youtu-GraphRAG: Vertically Unified Agents for Graph Retrieval-Augmented Complex Reasoning
- Youtu-GraphRAG introduces a vertically unified agentic paradigm for graph retrieval-augmented complex reasoning, integrating a Seed Graph Schema (defines entity/relation/attribute types), an Extraction Agent (schema-guided knowledge extraction), Dually-Perceived Community Detection (fuses topology and semantics), a Four-Level Knowledge Tree (hierarchical knowledge organization), an Agentic Retriever (schema-aligned query decomposition), a Planning Component (decomposes complex queries), a Reflection Component (iteratively refines reasoning), Historical Memory (stores agent's reasoning/retrieval), and an LLM (performs various language tasks).
- This framework jointly optimizes graph construction and retrieval by bounding both processes with a dynamically expanding graph schema, enabling robust and generalizable reasoning across different knowledge granularities.
- The framework significantly improves cost-effectiveness and accuracy by reducing token consumption and enhancing multi-hop reasoning, demonstrating strong adaptability for seamless domain transfer.
- Memory-R1: introduces a reinforcement learning framework that enhances LLM agents with active memory management and utilization through a Memory Manager and an Answer Agent.
- The Memory Manager learns to perform structured Memory Operations (ADD, UPDATE, DELETE, NOOP) on an External Memory Bank, while the Answer Agent applies Memory Distillation to filter and reason over retrieved memories.
- Both agents are fine-tuned using PPO or GRPO, enabling adaptive memory management and use with minimal supervision and achieving strong performance on multi-session dialogue tasks.
Survey of Specialized Large Language Model
- Specialized Large Language Models: introduces a comprehensive survey examining the progression of specialized LLMs from early domain adaptation to sophisticated native architectures across healthcare, finance, legal, and technical domains.
- The survey systematically analyzes architectural innovations, application successes, and persistent challenges, identifying key technological trends and performance characteristics of 48 cutting-edge models developed between 2022-2025.
- It highlights how innovations in dataset, training architecture, evaluation standards, retrieval augmentation, tool use, and memory address fundamental limitations of general-purpose LLMs in professional applications, consistently yielding performance gains on domain-specific benchmarks.
SwizzlePerf: Hardware-Aware LLMs for GPU Kernel Performance Optimization
- SwizzlePerf: introduces a hardware-aware LLM workflow that automatically generates spatial optimizations for GPU kernels by integrating parsed context, LLM code generation, and a bottleneck history buffer for iterative refinement.
- The framework leverages workload-specific memory access patterns, architecture specifications, and profiling logs to enable LLMs to tailor software-level optimizations to the underlying hardware.
- By imitating human performance engineers, SwizzlePerf autonomously finds optimal swizzling patterns for GPU kernels in minutes, significantly improving L2 hit rates and achieving substantial speedups.
Validating Generative Agent-Based Models for Logistics and Supply Chain Management Research
- GABM Validation Framework: introduces a dual-validation framework for Generative Agent-Based Models (GABMs) that assesses LLM-powered agents' fidelity to human behavior, including surface-level behavioral equivalence testing and process-level decision validation.
- The framework utilizes Two One-Sided Tests (TOST) for surface-level validation to compare GABM outputs with human behavioral baselines, and Structural Equation Modeling (SEM) for process-level validation to examine underlying decision-making pathways.
- This multi-level approach addresses the challenge that AI models can achieve output equivalence without replicating authentic human decision processes, providing systematic standards for rigorous GABM development and responsible LLM adoption in Logistics and Supply Chain Management (LSCM).
Symphony: A Decentralized Multi-Agent Framework for Scalable Collective Intelligence
- Symphony: introduces a decentralized multi-agent system, with a decentralized ledger (records capabilities), a Beacon-selection protocol (dynamic task allocation), weighted result voting (aggregates CoT results), Worker Nodes (host LLMs), Local Engine (quantized LLM), Stage-specific prompts (contextual instructions), Communicator (secure messaging), Gateways (standardized APIs), Planning Agents (decompose tasks), and Execution Agents (execute sub-tasks), enabling lightweight LLMs on edge devices to coordinate for scalable collective intelligence.
- This framework addresses challenges of centralized orchestration by providing a privacy-saving, scalable, and fault-tolerant design with low overhead, allowing efficient task allocation and robust operation across heterogeneous devices.
- Symphony demonstrates superior performance on reasoning benchmarks, achieving significant accuracy gains and robustness across models, while lowering hardware requirements and fostering decentralized agent economies.
A Symbolic Adversarial Learning Framework for Evolving Fake News Generation and Detection
- SALF (Symbolic Adversarial Learning Framework): introduces a novel framework for evolving fake news generation and detection, with a generator agent crafting deceptive narratives and a detection agent identifying flaws through structured debates, both iteratively refining their strategies via agent symbolic learning.
- The framework leverages LLMs to define learnable weights as agent prompts and simulates back-propagation and gradient descent using natural language representations, enabling adaptive and interpretable adversarial training.
- SALF demonstrates effectiveness by generating sophisticated fake news that degrades state-of-the-art detection performance and simultaneously refines detectors to improve their ability to identify refined content.
Instructional Agents: LLM Agents on Automated Course Material Generation for Teaching Faculties
- Instructional Agents: introduces a multi-agent LLM framework for automated course material generation, simulating role-based collaboration among Teaching Faculty, Instructional Designer, Teaching Assistant, Course Coordinator, Program Chair, and Test Student agents, guided by the Analyze, Design, and Develop phases of the ADDIE instructional design framework.
- The framework produces cohesive and pedagogically aligned instructional materials, including learning objectives, syllabi, LaTeX-based slides, slide scripts, and assessments, and operates in four modes: Autonomous, Catalog-Guided, Feedback-Guided, and Full Co-Pilot, to balance automation and human involvement.
- Instructional Agents aims to reduce educator workload, support content standardization, and enable scalable curriculum development, particularly for under-resourced institutions, by integrating human oversight and pre-existing data.
- RLTR (Reinforcement Learning with Tool-use Rewards): introduces a novel framework that decouples LLM agent training by focusing on single-objective optimization of the Planner (core planning component) using a reward signal based on tool-use completeness, thereby improving action planning and overall response quality.
- The framework addresses challenges of imbalanced optimization and scarce verifiable data by employing a Comp. Checker (Verification LLM) to evaluate tool invocation sequences, which is more reliable than assessing final response content.
- The Planner is initialized via Cold Start (knowledge distillation and rejection sampling) and then optimized through Multi-Turn RL, with the optimized Planner subsequently paired with a Summarizer (LLM) to generate the final end-to-end response.
Democracy-in-Silico: Institutional Design as Alignment in AI-Governed Polities
- Democracy-in-Silico: introduces an agent-based simulation where LLM Agents, embodying Complex Personas, govern themselves through a Legislative Cycle under various Institutional Design rules and Stressors, with a Deliberation Engine managing interactions, and Simulation Logs feeding into Measurement, including the Power-Preservation Index, Constitutional AI Charter, and an AI Mediator, to explore institutional design as an AI alignment mechanism.
- The framework tasks LLMs to embody agents with traumatic memories, hidden agendas, and psychological triggers, engaging in deliberation, legislation, and elections under stressors like budget crises and resource scarcity.
- The simulation demonstrates that institutional design, specifically a Constitutional AI charter and a mediated deliberation protocol, significantly reduces corrupt power-seeking behavior and enhances citizen welfare.
Can LLMs Generate Behaviors for Embodied Virtual Agents Based on Personality Traits?
- Embodied Virtual Agent System: introduces a framework that leverages personality prompting with LLMs to generate verbal and non-verbal behaviors for virtual agents, utilizing a Prompt, LLM, Personality Context, Non-Verbal Action List, Non-verbal Animation Description Generation Module, Animation Clips, and an Embodied Virtual Agent System with dedicated control modules.
- The system's pipeline generates verbal responses and selects appropriate nonverbal actions from a predefined list, ensuring alignment with the intended personality traits.
- It unifies LLM-generated speech with corresponding nonverbal actions, including facial expressions, body gestures, and voice characteristics, for coherent and personality-aligned virtual agent behaviors.
Learning Game-Playing Agents with Generative Code Optimization
- Trace framework: introduces an LLM-based generative optimization approach for learning game-playing agents, featuring an LLM Optimizer (OptoPrime) that refines a Policy (Python Program) using Trace Module, Trace Bundle, Trace Optimizer, Object-Centric Atari Environments (OCAtari), Execution Traces, Staged Feedback, and Policy Parameters.
- The approach treats decision-making policies as self-evolving Python code, enabling agents to self-improve through execution traces and natural language feedback with minimal human intervention.
- This method achieves competitive performance with deep reinforcement learning baselines in Atari games, using significantly less training time and fewer environment interactions, while maintaining interpretable and human-readable policies.
Aegis: Taxonomy and Optimizations for Overcoming Agent-Environment Failures in LLM Agents
- Aegis: introduces a framework for optimizing system environments to improve LLM agent reliability, featuring environment observability enhancement, common computation offloading, and speculative agentic actions.
- This approach addresses agent-environment interaction failures by enhancing information gathering, offloading deterministic reasoning, and reducing resource consumption through preemptive actions.
- The framework significantly improves task success rates and reduces monetary costs by making the environment more supportive and efficient for LLM agents, without modifying the agents themselves.
Multi-Agent Reinforcement Learning in Intelligent Transportation Systems: A Comprehensive Survey
- Multi-Agent Reinforcement Learning (MARL): introduces a comprehensive survey of MARL applications in Intelligent Transportation Systems, categorizing approaches by coordination models and learning algorithms, including value-based, policy-based, and actor-critic methods.
- The survey details MARL applications across key ITS domains, reviews common simulation platforms and benchmarks, and identifies core challenges like scalability and the sim-to-real transfer gap.
- Future research directions emphasize federated learning, safety-aware policy design, robust communication protocols, and integration with edge computing to advance practical and scalable ITS solutions.
Regulation-Aware Game-Theoretic Motion Planning for Autonomous Racing
- RA-GTP (Regulation-Aware Game-Theoretic Planner): introduces a regulation-aware motion planning framework for autonomous racing, with all RC-MPC, MLD framework, MLD Right-of-Way Constraints, MLD Collision Avoidance Constraints, MLD Sample-and-Hold Dynamics, GNEP, IBR scheme, and Regulation-Constrained Racing Game (G) components, where the attacker reasons over the defender's regulation-constrained behavior to generate safe and non-conservative overtaking strategies.
- The framework models vehicle interactions as a non-cooperative, two-player, finite-horizon differential game, formalizing it as a Generalized Nash Equilibrium Problem (GNEP) and approximating its solution using an Iterative Best Response (IBR) scheme.
- Each agent solves a Regulation-Compliant Model Predictive Control (RC-MPC) problem, where racing rules like right-of-way and collision avoidance responsibilities are encoded using Mixed Logical Dynamical (MLD) constraints.
- CODA (Coordinating the Cerebrum and Cerebellum for a Dual-Brain Computer Use Agent with Decoupled Reinforcement Learning): introduces a novel trainable compositional framework that synergizes a Planner (high-level thought generation) with an Executor (concrete GUI action execution), trained via a two-stage pipeline using Reward Signal (training feedback calculation) and Decoupled RL (Planner-focused reinforcement learning) to process User Instruction (task definition input) and generate Action (GUI command output).
- The training pipeline leverages a Task Generator (high-level task creation) and Judge System (reward signal generation) within a Distributed VM System (parallel task execution) to collect diverse Trajectories (agent interaction data) for both specialized and generalized Planner training stages.
- This decoupled approach, inspired by the human brain's cerebrum and cerebellum, enables the Planner to adapt through experience while the Executor provides stable, software-agnostic GUI grounding, addressing the trade-off between generalist planning and precise execution in GUI automation.
Evaluating Language Model Reasoning about Confidential Information
- PasswordEval benchmark: introduces "Evaluating Language Model Reasoning about Confidential Information", with PasswordEval benchmark (evaluates contextual robustness), Language Model (under test), User Prompt (user input/request), System Prompt (defines rules/context), Confidential Information (data to protect), Password (access credential), Evaluation Criteria (metrics for performance), Data Generation Pipeline (creates scenarios), Multi-turn Setting (multiple password verification), Adversarial Jailbreaks (stress-testing strategies), and Reasoning Traces (internal LLM thought process), where the paper evaluates LLMs' ability to handle confidential information under various conditions, including adversarial pressure and multi-turn interactions.
- The benchmark measures contextual robustness by tasking LLMs to conditionally reveal confidential information only when the correct password is provided, using metrics like CompliantAcc, NonCompliantAcc, ConfInfoLeak, and PasswordLeak.
- PasswordEval reveals that current LLMs struggle with this task, often leaking confidential information through reasoning traces, and that reasoning capabilities do not consistently improve rule-following, highlighting security concerns for high-stakes deployments.
- InquireMobile (VLM-based Mobile Agent to Request Human Assistance via Reinforcement Fine-Tuning): introduces a novel model designed to teach VLM-based mobile agents to request human assistance through reinforcement fine-tuning, which includes a Vision Encoder (perceives visual input), an LLM (processes instructions/reasons), Supervised Fine-tuning (SFT) (acquires structured outputs), Group Relative Policy Optimization (GRPO) (enhances reasoning/inquiry), Rule-based Action-level Reward (guides GRPO training), and an Interactive Pre-action Reasoning Mechanism (proactively inquires user).
- The model employs a two-stage training strategy, starting with SFT for robust format acquisition and followed by GRPO training to enhance reasoning and thinking capabilities, achieving a 46.8% improvement in inquiry success rate.
- The paper also introduces InquireBench, a comprehensive benchmark designed to evaluate mobile agents' capabilities in safe interaction and proactive inquiry with users, demonstrating the necessity of proactive user engagement in agent-driven automation.
CompLex: Music Theory Lexicon Constructed by Autonomous Agents for Automatic Music Generation
- LexConstructor: introduces an automatic music lexicon construction model that generates CompLex, a comprehensive music theory lexicon, using a multi-agent algorithm composed of Category Architect, Item Builder, Property Designer, Supervisor Agent, and Value Explorer Agents, leveraging a Reference MIDI Dataset and LLMs.
- This multi-agent algorithm operates in two stages, Lexicon Outline Creation and Lexicon Content Generation, to determine the lexicon's structure and populate it with property-value pairs, while automatically detecting and mitigating hallucinations through a Question-Answering communication strategy.
- The framework significantly reduces manual effort in music lexicon development and enhances text-to-music generation models by providing structured music theory knowledge, improving completeness, accuracy, non-redundancy, and executability.
Private, Verifiable, and Auditable AI Systems
- End-to-End Secure and Auditable AI System: introduces a technical framework for building trustworthy AI systems by integrating cryptographic and secure computing techniques across the AI supply chain, including zkSNARKs (verifiable computation proofs), TEEs (secure hardware enclaves), MPC (distributed private computation), and authenticated delegation protocols (AI agent permissions), to address privacy, verifiability, and auditability challenges in foundation model-based AI.
- The framework leverages zkSNARKs for verifiable ML evaluation and data attestations, enabling proofs of model performance and data provenance without revealing sensitive information.
- It also proposes Private Retrieval Augmented Generation (PRAG) for secure, private querying of distributed databases, and integrates personhood credentials to verify human users behind AI agents, enhancing trust and accountability.
SWIRL: A STAGED WORKFLOW FOR INTERLEAVED REINFORCEMENT LEARNING IN Mobile GUI CONTROL
- SWIRL (Staged Workflow for Interleaved Reinforcement Learning): introduces a multi-agent training framework for mobile GUI control, featuring a Navigator (high-level planner, LLI generator) and an Interactor (low-level executor, atomic actions), which leverages interleaved reinforcement learning for stable and efficient coordination.
- The framework decomposes multi-agent training into a warm-up initialization (independent agent pre-training) and an interleaved update (alternating agent optimization) stage, ensuring robust optimization and O(1) actor memory usage.
- SWIRL further incorporates an online reweighting mechanism (filters low-quality samples) to enhance training stability and accelerate convergence, demonstrating superior performance in GUI control and mathematical reasoning tasks.
PG-Agent: An Agent Powered by Page Graph
- PG-Agent (An Agent Powered by Page Graph): introduces a framework that transforms sequential GUI operation records into a structured Page Graph, which serves as a GUI prior knowledge base, and leverages a RAG pipeline to retrieve Guidelines for a Multi-Agent Group to enhance GUI navigation.
- The framework's Reconstruction Pipeline converts Episodes into a Page Graph through Page Jump Determination, Node Similarity Check, and Page Graph Update, explicitly modeling complex page transition relationships.
- The Multi-Agent Group, comprising Observation, Global Planning, Sub-Task Planning, and Decision Agents, utilizes the RAG Pipeline's retrieved Guidelines from the Page Graph to achieve targeted planning and robust action selection in unseen GUI scenarios.
InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
- InternVL3.5: introduces a new family of open-source multimodal models, incorporating Cascade RL (two-stage reasoning enhancement), ViR (dynamic visual token resolution adjustment), and DvD (decoupled vision-language model deployment) to advance versatility, reasoning, and efficiency.
- The Cascade RL framework, with its offline and online stages, significantly improves reasoning capabilities, while ViR and DvD optimize inference speed and computational load balancing.
- InternVL3.5-Flash, an efficient variant, integrates ViR through Visual Consistency Learning, enabling dynamic resolution adjustment without compromising performance, and achieves state-of-the-art results across various multimodal tasks.
BUILDING SELF-EVOLVING AGENTS VIA EXPERIENCE-DRIVEN LIFELONG LEARNING: A FRAMEWORK AND BENCHMARK
- ELL (Experience-driven Lifelong Learning): introduces a framework for building self-evolving agents capable of continuous growth through real-world interaction, featuring Perception, Memory, Learning, Reasoning, and Action modules.
- The framework is supported by StuLife, a benchmark simulating a student's college journey to evaluate lifelong learning capabilities, including memory retention, skill transfer, and self-motivated behavior.
- The research reveals current LLMs' limitations in self-motivation and long-term memory, emphasizing context engineering's crucial role in advancing AGI.
Optimizing Highway Traffic Flow in Mixed Autonomy: A Multiagent Truncated Rollout Approach
- Multiagent Truncated Rollout Approach: introduces a novel method for optimizing highway traffic flow in mixed autonomy, integrating a PDE-ODE coupled model, a system-level density evolution equation, and a distributed coordination control framework.
- The approach employs independent MPC controllers for each CAV, an agent-by-agent sequential optimization mechanism for explicit cooperation, and a truncated rollout scheme to adaptively shorten the optimization horizon based on objective function bounds.
- This framework enhances CAV speed coordination, improves highway throughput, and reduces computational overhead by leveraging real-time policy sharing and dynamic horizon adjustment, ensuring system stability and performance improvement.
Real-Time Model Checking for Closed-Loop Robot Reactive Planning
- Agent Architecture: introduces a novel real-time model checking approach for closed-loop robot reactive planning, with Robot (mobile platform), LIDAR (2D laser scanner), Motors (actuation), Raspberry Pi 3 Model B (onboard computer), Environment (robot's surroundings), Disturbance D (environmental obstacle), Task Controller (orchestrates tasks), Model Checking (planning algorithm), Tasks (closed-loop control systems: Default/Finite straight/Rotate left/Rotate right), Disturbance-Focused Transition System (robot behavior model), Nondeterministic Finite Automaton (LTL property checker), Product Transition System (combined state-space model), Lateral Partitions (spatial reasoning for turns), Longitudinal Partitions (spatial reasoning for straight paths), Safe Zone (collision-free region), and Shield Partition (proximal disturbance detection), where it enables efficient multi-step planning and obstacle avoidance on a low-powered autonomous robot.
- This framework generates plans in situ based on "core" knowledge and attention, chaining temporary control systems to counteract disturbances without relying on pre-computed data or extensive prior experience.
- The approach utilizes a novel discretization of 2D LiDAR data and forward depth-first search to create efficient multi-step plans for local obstacle avoidance, demonstrating improved performance over single-step reactive agents in cul-de-sac and playground scenarios.
SecureV2X: An Efficient and Privacy-Preserving System for Vehicle-to-Everything (V2X) Applications
- SecureV2X: introduces an efficient and privacy-preserving system for Vehicle-to-Everything (V2X) applications, with CryptoDrowsy (Secure driver drowsiness detection module), FastSec-YOLO (Secure red-light violation detection module), Client (Vehicle/user holding EEG or image data), Server (Edge server/cloud holding model weights), Secure Mediating Agent (Third-party for Beaver's triples distribution), CrypTen MPC Framework (Underlying secure computation library), Private Model Weights (Proprietary neural network parameters), Private Data (Sensitive user input, e.g., EEG, video), Secure Computation (Joint execution of inference protocols), Secure Inference Setting (Operational environment for secure V2X applications), and Violation Alert! (Output for detected red-light violations), which enables secure neural network inferences between servers and vehicles for critical safety tasks.
- The system addresses privacy concerns in V2X by implementing two multi-agent applications: secure drowsiness detection using CompactCNN and secure red-light violation detection via YOLOv5, both built upon novel cryptographic protocol constructions.
- SecureV2X significantly outperforms state-of-the-art secure systems in terms of inference speed, communication rounds, and computational efficiency, making it suitable for real-time, time-sensitive safety applications while preserving user data privacy and model security.
A Concurrent Modular Agent: Framework for Autonomous LLM Agents
- CMA (Concurrent Modular Agent): introduces a framework orchestrating multiple asynchronous LLM-based modules, a shared vector store, and inter-module communication for coherent, fault-tolerant agent behavior.
- This framework enables flexible, adaptive, and context-dependent behavior by offloading reasoning to LLMs and allowing intention to emerge from language-mediated interactions among autonomous processes.
- Demonstrated on physical robotic platforms (Plantbot, ALTER3), the architecture supports robust, scalable AI systems exhibiting emergent cognitive phenomena like self-awareness and identity formation.
STARec: An Efficient Agent Framework for Recommender Systems via Autonomous Deliberate Reasoning
- STARec (Slow-Thinking Augmented agent framework): introduces an LLM-based agent framework for recommender systems, featuring a STARec Agent (main processing unit) with a Memory Module (stores user preferences), Fast Thinking for Personalized Ranking (intuitive item ranking), and Slow Thinking for Memory Update (deliberate preference refinement), all supported by Anchored Reinforcement Training (two-stage learning paradigm) comprising SFT Anchoring (foundational capability instillation) with a Teacher Model (generates reasoning data) and Filter and Augment (refines SFT dataset), and RL Enhancing (policy optimization) with a GRPO Algorithm (reinforcement learning optimizer) and Ranking-Oriented Reward (guides ranking decisions), integrated through a Continuous Learning Cycle (dynamic adaptation mechanism).
- This framework models each user as an autonomous agent with dual-process cognition, enabling both rapid, intuitive responses for immediate interactions and slow, deliberative reasoning for continuous preference adaptation and memory refinement.
- The anchored reinforcement training strategy bridges the gap between LLMs' generic knowledge and domain-specific reasoning, using structured knowledge distillation and preference-aligned reward shaping to cultivate intrinsic slow thinking and dynamic policy adaptation.
Governance-as-a-Service: A Multi-Agent Framework for AI System Compliance and Policy Enforcement
- GaaS (Governance-as-a-Service): introduces a modular, policy-driven enforcement layer for AI systems, with Autonomous Agents (LLM-based, rule-based), LLM Agent, Finance Bot, Infrastructure Agent, Policy Loader, Policy Engine, Trust Computation, Violation Checker, Enforcement Engine, Audit Logger, Trust Registry, Secure Release Gate, Compliance Pipeline, Downstream Systems, End Users / Markets, and Human Oversight, designed to govern agent outputs at runtime without modifying internal model logic.
- This framework operates through declarative rule sets and a Trust Factor mechanism, scoring agents based on longitudinal compliance and severity-aware violation history to support coercive, normative, and adaptive interventions.
- GaaS aims to provide scalable, auditable, and adaptive AI oversight for decentralized, open-source agentic ecosystems by treating governance as a provisioned runtime service.
- Agentic AI: introduces a comprehensive framework for edge general intelligence, with Perception (acquires multimodal data), Memory (stores, retrieves knowledge), Reasoning (plans, reasons, decides), and Action (executes decisions, interacts) modules, enabling autonomous perception-reasoning-action loops in dynamic edge environments.
- This framework leverages LLMs as cognitive cores for semantic comprehension and planning, integrates external tools/APIs to extend capabilities, and utilizes a continuous feedback loop for iterative self-refinement and adaptation.
- The system aims to overcome limitations of traditional edge AI by providing robust, scalable, and human-aligned solutions for complex tasks in resource-constrained 6G-enabled networks.
Bias Mitigation Agent: Optimizing Source Selection for Fair and Balanced Knowledge Retrieval
- Bias Mitigation Agent: introduces a supervisor-based multi-agent system for bias mitigation, with a Manager Agent (coordinates workflow), Knowledge Agent (retrieves documents), Bias Detector Agent (evaluates bias), Source Selector Agent (selects unbiased sources), and Writer Agent (synthesizes answer), where the system optimizes source selection for fair and balanced knowledge retrieval.
- This framework uses a centralized Manager Agent to supervise execution flow, maintain system state, and coordinate decisions among specialized Worker Agents (Knowledge, Bias Detector, Source Selector, Writer) to ensure relevant and minimally biased content.
- The system supports "No Source Selection", "Zero-Shot", and "Few-Shot" operational modes, allowing flexible trade-offs between computational efficiency, fairness enforcement, and generalization capabilities in knowledge retrieval tasks.
FALCON: Autonomous Cyber Threat Intelligence Mining with LLMs for IDS Rule Generation
- FALCON (Autonomous Cyber Threat Intelligence Mining with LLMs for IDS Rule Generation): introduces an autonomous agentic framework that generates deployable Intrusion Detection System (IDS) rules from Cyber Threat Intelligence (CTI) data, incorporating LLM-driven generation, multi-phased validation, and human oversight to automate the entire rule-generation pipeline.
- The framework addresses the challenge of rapidly evolving cyber threats by enabling real-time IDS rule generation and updates for both network (Snort) and host-based (YARA) environments, ensuring syntactic correctness, semantic alignment, and performance optimization.
- FALCON integrates LLM-driven data mining with iterative feedback loops and human oversight, significantly reducing manual effort and enhancing the agility and accuracy of threat detection systems.
MUA-RL: MULTI-TURN USER-INTERACTING AGENT REINFORCEMENT LEARNING FOR AGENTIC TOOL USE
- MUA-RL (Multi-turn User-interacting Agent Reinforcement Learning for agentic tool use): introduces a novel reinforcement learning framework that integrates LLM-simulated users into the RL loop for agentic tool use, including an Agent LLM, User LLM, Tool LLM/MCP server, External Database, Reinforcement Learning Loop, GRPO, Synthesized Database, Trajectory Verifiers, Reward Mechanism, Cold-start Training Phase, and Multi-turn Rollout Process.
- This framework enables autonomous learning for agents to efficiently communicate with users and utilize various tools to solve dynamic multi-turn interaction problems.
- MUA-RL employs a simplified, task-oriented reward design and a cold-start phase to develop robust behavioral patterns and enhance generalization across diverse tool-using tasks.
Bimodal Dynamics of the Artificial Limit Order Book Stock Exchange with Autonomous Traders
- ASME (Artificial Stock Market Exchange): introduces a framework for an artificial stock market with autonomous, myopic traders interacting through a limit order book, revealing intrinsic bistability and complex dynamics.
- The framework utilizes an HMM to analyze bifurcative dynamics, identifying two distinct long-run price equilibria: a deterministic zero-price state and a persistent positive-price equilibrium.
- The paper employs Logistic Regression and Gradient Boosting Machines to predict trajectory outcomes and various complexity measures (Fractal Dimension, Entropy, LLE) to characterize the system's structured, yet dynamically rich, behavior.
- MATRIX (Multi-Agent simulaTion fRamework for safe Interactions and contexTual clinical conversational evaluation): introduces a structured, extensible framework for safety-oriented evaluation of clinical dialogue agents, comprising a Structured Safety Library, PatBot (LLM-based Simulated Patient Agent), a Clinical History Taking Agent (LLM Target System), BehvJudge (LLM-based Safety Evaluator), a Clinical Use-Case Specific Context, and System Performance Output.
- The framework enables systematic and scalable safety evaluation by unifying structured safety engineering with validated conversational AI evaluation, supporting regulator-aligned safety auditing.
- It benchmarks LLM agents across simulated clinical dialogues, identifying failure patterns in safety-critical scenarios, and demonstrates that LLM-based evaluators can surpass human performance in hazard detection.
- DELIVER (Directed Execution of Language-instructed Item Via Engineered Relay): introduces a fully integrated system for cooperative multi-robot pickup and delivery, with Natural Language Understanding (parses natural language commands), Voronoi Partitioning (divides environment into robot regions), Pickup and Drop Agent Identification (assigns robots to task endpoints), Active Agent Selection (selects robots for relay path), Relay Point Selection (calculates handover locations), and Relay Execution (manages robot movement and handoffs).
- The system unifies LLM-based natural language understanding, Voronoi-based spatial decomposition for region-aware planning, relay-point computation for inter-agent coordination, and execution through local finite-state machines with lightweight signaling.
- DELIVER demonstrates scalability and efficient agent utilization by reducing per-agent workload by up to 55% compared to single-agent systems, maintaining consistent mission cost and low coordination overhead.
Reasoning LLMs in the Medical Domain: A Literature Survey
- Reasoning LLMs in the Medical Domain: introduces a comprehensive literature survey on the current state and future potential of reasoning LLMs within the medical domain, examining their transformative role in healthcare applications.
- The survey analyzes enabling technological foundations like Chain-of-Thought and Reinforcement Learning, alongside emerging paradigms such as specialized medical LLMs, multi-agent systems, and innovative prompting architectures.
- It critically assesses current evaluation methodologies, addresses persistent challenges, and delineates a roadmap for developing reliable, safe, and ethically aligned LLMs for medical use.
Trustworthy Agents for Electronic Health Records through Confidence Estimation
- TrustEHRAgent: introduces a confidence-aware clinical agent for Electronic Health Records (EHR) that integrates step-wise confidence estimation (tracks uncertainty per step) and a confidence estimator (computes final confidence) to make threshold-based decision making (decides answer or reject) for clinical question answering.
- The framework leverages token probability (confidence score input) and weighted average (calculates final confidence) within its Confidence Estimator to derive a final confidence score, which is then compared against a predefined reliability threshold (τ) to either provide an answer (provides confident answers) or reject the query (abstains from uncertain queries).
- This approach enhances reliability by enabling the agent to transparently express uncertainty and abstain from answering when confidence is low, thereby preventing potential errors and improving patient safety in high-stakes medical contexts.
HIPLAN: Hierarchical Planning for LLM Agents with Adaptive Global-Local Guidance
- HIPLAN (Hierarchical Planning for LLM Agents with Adaptive Global-Local Guidance): introduces a hierarchical planning framework that provides adaptive global-local guidance to boost LLM-based agents' decision-making, with all components including LLM (generates milestones, hints, actions), Milestone Library (stores structured expert experience), Milestone Action Guide (provides global task direction), Step-Wise Hints (offers local action feedback), Expert Demonstrations (source for experience library), Milestones Extraction (segments trajectories into subgoals), Task-Level Similarity Search (retrieves relevant tasks), Milestone-Level Similarity Search (retrieves relevant trajectory fragments), Agent Policy (integrates guidance for actions), and Embeddings (vector representations for retrieval), enabling LLM-based agents to tackle complex, long-horizon tasks through integrated global and local guidance.
- The framework constructs a milestone library offline from expert demonstrations, which is then used during execution to retrieve relevant task and milestone-level experiences for generating dynamic global milestone action guides and local step-wise hints.
- This dual-level guidance mechanism enhances efficiency, controllability, and overall robustness by maintaining global coherence while adapting actions to dynamic local contexts, outperforming baselines on ALFWorld and WebShop benchmarks.
MovieCORE: COgnitive REasoning in Movies
- MovieCORE (COgnitive REasoning in Movies): introduces a novel video question answering (VQA) dataset designed to probe deeper cognitive understanding of movie content, generated using an agentic brainstorming approach.
- This approach leverages multiple LLMs as specialized agents—including a Critic Agent (MC), System II VQA Expert, Skeptical Researcher, Detective, and Meta Reviewer—to generate and refine high-quality, thought-provoking question-answer pairs, validated by Human Reviewers and informed by Video Context Extraction (MiniCPM-v2.6).
- The paper also proposes Agentic Choice Enhancement (ACE), a post-training plugin that improves existing VLMs' reasoning capabilities by using an ACE Existing VLM, ACE Beam Search, and ACE Llama-3.2 for response generation and re-ranking.
- GitTaskBench: introduces a benchmark for code agents, evaluating their ability to solve real-world tasks by leveraging code repositories, which includes Task & Repository Selection, Completeness Verification, an Execution Framework for agent workflow, and an Evaluation Framework with defined success criteria and a practical utility (alpha-value) metric.
- This benchmark systematically assesses agents' overall coding mastery, task-oriented execution, and autonomous environment provisioning across 54 real-life, multimodal tasks from 7 domains, using human-curated evaluation scripts.
- It also proposes a novel "alpha-value" metric to quantitatively assess agent economic benefits, integrating task success, token cost, and average developer salaries, providing actionable insights for agent deployment.
Interactive Evaluation of Large Language Models for Multi-Requirement Software Engineering Tasks
- Interactive, Dependency-Grounded Assessment: introduces a novel interactive evaluation framework for LLMs on multi-requirement programming tasks, featuring a structured, feedback-driven dialogue between an Interviewer (LLM-based, generates feedback) and an Interviewee (LLM under evaluation), supported by Task specification (defines problem parameters), Reference Solution (ground-truth for guidance), Evaluation Guidelines (criteria for assessment), History (stores interaction dialogue), Report (structured performance analysis), Executor (runs interviewee code), Solution Output (results from code execution), Solution (interviewee's code response), Solution Protocol (defines solution structure), and Delivery Format (specifies output format).
- This framework models tasks as requirement dependency graphs, allowing an LLM-based interviewer to provide minimal, targeted hints to an interviewee model for error correction and constraint fulfillment.
- The dynamic protocol enables fine-grained diagnostic insights into model behavior, uncovering strengths and systematic weaknesses that static benchmarks fail to measure, and guides the interviewee through iterative refinement loops.
Judicial Requirements for Generative AI in Legal Reasoning
- No single overarching framework is proposed; the paper analyzes existing AI enhancement mechanisms: introduces an analysis of AI enhancement mechanisms, including Fine-tuning, Retrieval-Augmented Generation (RAG), Task Decomposition and Chained Prompts, Tree of Thoughts (ToT), Neuro-Symbolic AI, Multi-Agent Systems, Structured Self-Evaluation, and Logit-based Confidence Scoring, to assess their potential in meeting judicial requirements for generative AI in legal reasoning.
- The study uses the IRAC (Issue-Rule-Application-Conclusion) model as an analytical framework, focusing on the challenging phases of legal adjudication: determining the applicable Rule (R) and performing the Application (A) of that rule to the facts of a case.
- The findings indicate that while these techniques can address specific challenges, significant challenges remain, particularly in tasks requiring discretion and transparent, justifiable reasoning, concluding that the most effective current role for AI in law is a dual one: as a high-volume assistant for simple, repetitive cases and as a sophisticated "sparring partner" for human experts in complex matters.
A Survey on Cloud-Edge-Terminal Collaborative Intelligence in AIoT Networks
- CETCI (Cloud-Edge-Terminal Collaborative Intelligence): introduces a comprehensive survey on cloud-edge-terminal collaborative intelligence in AIoT networks, with Cloud Layer (centralized computing, global storage), Edge Layer (distributed processing, real-time inference), Terminal Layer (data acquisition, IoT device control), Network Virtualization (flexible network infrastructure), Container Orchestration (application deployment management), Software-Defined Networking (SDN) (centralized network control), AI/ML Integration Platforms (intelligent decision-making), Resource Management (optimizes task offloading, allocation), Task Offloading (learning-based, game theory/optimization), Resource Allocation (learning-based, energy-aware, QoS-driven), Optimization Techniques (linear/convex programming, game theory), Collaborative Learning (develops intelligent models), Federated Learning (FL) (privacy-preserving, robust learning), Distributed Deep Learning (DDL) (model/data parallelism), Model Evolution (compression, distillation, incremental learning), RL Optimization (resource management, multi-agent RL), Security & Privacy (protects data flow, system integrity), Security Threats (data breaches, DoS attacks), Security Mechanisms (encryption, authentication, IDS/IPS), Privacy Technologies (FL, differential privacy, homomorphic encryption), Data Management & Communication (foundational data infrastructure), Data Acquisition & Preprocessing (filtering, aggregation, compression), Storage & Retrieval (edge caching, distributed storage), Communication & Optimization (MQTT/CoAP, bandwidth optimization), Performance Metrics (latency, energy, utilization, QoS/QoE), and Application Domains (smart manufacturing, transportation, healthcare, cities, agriculture), where the paper systematically analyzes architectural components, enabling technologies, and collaboration paradigms across heterogeneous network infrastructures.
- The survey provides a tutorial-style review for beginners in CISAIoT, examining core technologies like network virtualization, container orchestration, and software-defined networking, while presenting multi-perspective categorizations of collaboration paradigms.
- It further explains intelligent collaboration learning frameworks by reviewing recent advances in federated learning, distributed deep learning, edge-cloud model evolution, and reinforcement learning-based approaches, discussing challenges and future development trends including LLMs and agents.
CausalMACE: Causality Empowered Multi-Agents in Minecraft Cooperative Tasks
- CausalMACE (Causality Empowered Multi-Agents in Minecraft Cooperative Tasks): introduces a holistic causality planning framework designed to enhance multi-agent systems in Minecraft, incorporating causality to manage dependencies among subtasks, with Judger (defines objectives/feedback), Planner (decomposes/graphs dependencies), Planner-Task Decomposition (breaks into subtasks), Planner-Factual Graph (FG) (initial dependency graph), Planner-Counterfactual Graph (CG) (causal inference graph), Planner-Graph Refinement (refines graph causally), Planner-ATE (Average Treatment Effect) (quantifies causal effect), Planner-LLMs (decompose/identify dependencies), Worker (assigns/executes subtasks), Worker-Agent Assignment (distributes subtasks), Worker-Path Sampling (explores execution paths), Worker-Busy Rate (br) (balances workload), Agents (execute/reflect autonomously), and Game Environment (Minecraft interactive world) components.
- The framework leverages an overarching task graph for global task planning and a causality-based module for dependency management, utilizing LLMs for task decomposition and causal intervention to refine the task graph.
- CausalMACE achieves state-of-the-art performance in multi-agent cooperative tasks by ensuring efficient task arrangement and execution through structured dependency management and balanced workload distribution.
VistaWise: Building Cost-Effective Agent with Cross-Modal Knowledge Graph for Minecraft
- VistaWise: introduces a cost-effective agent framework for Minecraft, integrating an LLM, text-modal and cross-modal graph construction, task-specific information retrieval, a memory stack, and a desktop-level skill library.
- The framework enhances decision-making by combining domain-specific knowledge from a cross-modal knowledge graph with real-time visual perception via a finetuned object detection model.
- VistaWise enables direct desktop control through mouse and keyboard inputs, reducing reliance on environmental APIs and achieving state-of-the-art performance in open-world tasks with significantly lower development costs.
- AppAgent-Pro: introduces a proactive GUI agent system that actively integrates multi-domain information based on user instructions, with its Comprehension Stage (analyzes user instructions), Cognitive Agent (LLM-based analysis/synthesis), Proactive Thinking (anticipates user needs), Execution Stage (autonomously interacts apps), Proactive Execution Agent (LLM-driven app interaction), Shallow Execution Mode (fast, surface-level retrieval), Deep Execution Mode (in-depth, iterative mining), Integration Stage (combines diverse information), and Personalization (leverages interaction history) components, designed to anticipate user needs and conduct in-depth multi-domain information mining.
- The system operates through a three-stage pipeline—Comprehension, Execution, and Integration—enabling it to proactively acquire relevant knowledge, understand user intent, perform appropriate actions, and integrate results into coherent outputs.
- AppAgent-Pro enhances efficiency, personalization, and depth of information access by moving beyond reactive LLM-based agents to a proactive paradigm that integrates and reasons across heterogeneous information domains.
Utilizing Training Data to Improve LLM Reasoning for Tabular Understanding
- LRTab (Learn then Retrieve): introduces a novel prompting-based reasoning approach that integrates training data insights by generating and retrieving "Prompt Conditions" to improve LLM tabular understanding.
- The framework leverages a Code-Augmented LLM to generate Chain-of-Thought responses and, for incorrect answers, employs a Prompt Condition Generation Module to predict and verify error-correcting conditions, which are then stored in a Knowledge Base.
- At inference, LRTab utilizes a Table Encoder and a Retrieval Module, refined by a Crossencoder Reranker, to retrieve the most relevant Prompt Conditions, providing additional context to the Code-Augmented LLM for accurate tabular reasoning.
Requirements Development and Formalization for Reliable Code Generation: A Multi-Agent Vision
- REDEFO (Requirements Development and Formalization): introduces a multi-agent framework for reliable code generation, with Analyst (interprets, structures NLRs), Formalizer (translates, assesses specifications), Coder (generates, verifies code), Knowledge Source (provides background knowledge), and Human Experts (provide review, feedback) components, designed to transform Natural Language Requirements (NLRs) into provably correct software artifacts through formal specification and verification.
- The framework leverages formal methods to bridge the gap between ambiguous NLRs and precise executable code, enabling rigorous reasoning, bug uncovering, and enforcement of critical properties throughout the software development process.
- REDEFO aims to enhance the quality and correctness of auto-generated software by integrating human-in-the-loop interaction and expert judgment at critical points, culminating in a trustworthy software product.
Bias-Adjusted LLM Agents for Human-Like Decision-Making via Behavioral Economics
- Persona-Based Approach: introduces a method for adjusting LLM biases using individual-level behavioral traits, with Behavioral Traits Dataset (stores individual human data), Persona Injection Module (assigns traits to LLMs), LLM Agents (simulate human decision-making), Economic Game Simulation Environment (executes decision-making scenarios), and Output Analysis Module (compares LLM output to human data), to better align LLM decisions with human behavior in economic games.
- This approach leverages the Econographics dataset to characterize LLM agents with unique personas based on real individual behavioral and demographic attributes, enabling the construction of a heterogeneous agent population.
- The method demonstrates improved alignment between simulated and empirical human behavior, particularly for the responder role in the ultimatum game, highlighting the potential of persona-conditioned LLMs for scalable human-like behavior simulation.
Reliable Weak-to-Strong Monitoring of LLM Agents
- MRT (Monitor Red Teaming) Workflow: introduces a standardized workflow for stress-testing and improving LLM agent oversight, featuring LLM Monitor Scaffolding, a Monitor LLM, LLM Agents, Threat Models, Evasion Strategies, SHADE-Arena, CUA-SHADE-Arena, and Human-in-the-Loop (HiLT) Monitoring, to evaluate monitoring systems against adversarial LLM agents.
- The workflow evaluates various monitor architectures, including a novel Hybrid Hierarchical-Sequential Scaffolding, demonstrating that architectural design is more impactful for reliability than the monitor's awareness level.
- The research also shows that targeted human oversight within the HiLT Monitoring component, particularly on pre-flagged suspicious activities, significantly improves true positive rates, enabling weaker models to reliably monitor stronger agents.
QAgent: An LLM-based Multi-Agent System for Autonomous OpenQASM programming
- QAgent: introduces an LLM-powered multi-agent system for autonomous OpenQASM programming, featuring a hierarchical design with specialized Dynamic-few-shot and Tools-augmented Coders that leverage RAG, few-shot learning, and CoT reasoning for code generation and refinement.
- The system dispatches tasks to either the Dynamic-few-shot Coder for structured problems or the Tools-augmented Coder for complex, parameterized tasks, employing iterative self-reflection for debugging and optimization.
- This multi-agent framework significantly enhances QASM code generation accuracy, making quantum programming more accessible and robust for non-experts.
- ARIS (Agreement-based Reflective Inference System): introduces a hybrid event extraction framework that systematically integrates a discriminative Sequence Tagger and a generative Self Mixture of Agents, leveraging model consensus, confidence-based filtering, and an LLM Reflection Module to resolve ambiguities.
- The framework employs Decomposed Instruction Fine-Tuning to equip the LLM with specialized capabilities for event subtasks, enhancing its accuracy and reliability in reflective reasoning.
- ARIS utilizes structured prompts, Triggers Reflection Prompt and Argument Reflection Prompt, to guide the Reflection Module (LLM) in classifying trigger candidates and validating argument roles for robust event extraction.
AT-CXR: Uncertainty-Aware Agentic Triage for Chest X-rays
- AT-CXR (Agentic Triage for Chest X-ray): introduces an uncertainty-aware agentic framework for chest X-ray triage, featuring Data Ingestion (detects/preprocesses cases), Uncertainty Check (computes confidence/OOD), Agentic Decision Routing (iteratively selects tools via guardrailed policy), and Triage and Explainability Artifacts (auto-sorts cases, generates CAM/LWI).
- The framework employs a Router, which can be either a deterministic rule-based or an LLM-decided policy, to select from a Toolbox of Diagnosis Assist (TTA, MoE, VLM), LWI Computation (segmentation, suppression, LWI), and Visualization (CAM) tools.
- This system aims to make autonomous, safe triage decisions under clinical constraints by estimating per-case confidence and distributional fit, enabling selective automation with auditable operation, and providing complementary operating points for throughput or accuracy.
Reliable Weak-to-Strong Monitoring of LLM Agents
- MRT (Monitor Red Teaming) Workflow: introduces a standardized workflow for stress-testing LLM agent monitoring systems, integrating LLM Agents, LLM Monitors with diverse Monitor Scaffolding (Baseline, Sequential, Hierarchical, Hybrid), Attackers using Evasion Strategies, Environments (SHADE-Arena, CUA-SHADE-Arena), a Human-in-the-Loop (HiLT) System, Tools, and Evaluation Metrics.
- The paper empirically evaluates monitor reliability under various threat models, agent/monitor awareness levels, and scaffolding designs, highlighting the hybrid scaffolding's superior robustness against adversarial attacks.
- The research demonstrates that architectural design (scaffolding) is more critical for improving monitor reliability than increased monitor awareness, enabling weaker models to effectively oversee stronger agents.
- Aleks (AI-powered Multi Agent System): introduces an AI-powered multi-agent system for autonomous scientific discovery, featuring a Domain Scientist Agent (provides domain knowledge/feedback), a Data Analyst Agent (proposes modeling strategies/refines analysis), a Machine Learning Engineer Agent (implements models/generates code/executes experiments), Shared Agent Memory (stores experimental records/facilitates communication), Episodic Memory (agent-specific task history), Semantic Memory (agent-specific knowledge base), Human Research Team (provides input/receives output), Research Questions & Datasets (initial input for discovery), and a Tool Space (MLE agent's execution environment), with provisions for Other Possible Agents (future specialized agents).
- Aleks autonomously conducts data-driven scientific discovery by iteratively formulating problems, exploring modeling strategies, and refining solutions without human intervention, leveraging specialized LLM-powered agents that collaborate through a shared memory architecture.
- The system balances automated exploration with interpretability and domain relevance, integrating domain knowledge and memory to achieve robust and coherent outcomes in scientific research, as demonstrated in a case study on grapevine red blotch disease.
DiscussLLM: Teaching Large Language Models When to Speak
- DiscussLLM: introduces a framework and dataset to teach LLMs the crucial skill of timely and valuable intervention in human conversations, with all its components, where it addresses the "When to Speak" problem by training models to proactively decide whether to remain silent or intervene with a helpful response.
- The framework utilizes a scalable two-stage data generation pipeline to synthesize a large-scale dataset of realistic multi-turn human discussions, each annotated with an intervention type and a conversational trigger.
- Two architectural baselines are explored: an integrated end-to-end generative model and a decoupled classifier-generator system, evaluating their ability to accurately time interventions and generate high-quality responses.
- The AI Data Scientist: introduces an autonomous LLM-powered agent that transforms raw data into actionable business recommendations, featuring a Data Cleaning Subagent (cleans, handles missing values, outliers), a Hypothesis Subagent (generates, tests data relationships), a Preprocessing Subagent (prepares data for modeling), a Feature Engineering Subagent (creates predictive features), a Model Training Subagent (trains predictive machine learning models), and a Call-To-Action Subagent (translates findings into recommendations).
- This framework emphasizes a hypothesis-driven approach, where specialized LLM Subagents work sequentially, passing structured metadata to ensure statistically validated insights guide each step from data preparation to final recommendations.
- The system automates the entire end-to-end data science workflow, enabling rapid generation of interpretable results and actionable strategies, significantly reducing the time from evidence to decision-making.
Teaching LLMs to Think Mathematically: A Critical Study of Decision-Making via Optimization
- Structured Roadmap for Advancing LLM Capabilities in Mathematical Programming: introduces a critical study of LLMs in mathematical optimization, proposing future directions via Structured Dataset Construction Framework (builds diverse, robust datasets), Modular Multi-Agent Architectures (decomposes tasks, assigns specialized LLMs), Chain of RAGs (iterative retrieval, external knowledge), Neuro-Symbolic Formulation (combines LLMs, symbolic solvers, verification), and Improved Prompting Strategies (adaptive, structured guidance), to enhance performance in complex optimization tasks.
- The roadmap addresses current LLM limitations in numerical reasoning, input length sensitivity, and reliance on surface-level pattern matching by integrating structured data, multi-agent collaboration, iterative knowledge retrieval, and formal verification.
- Key proposed components include a four-part dataset structure for capturing reasoning steps, specialized LLMs for subtasks, iterative RAG for dynamic context refinement, and neuro-symbolic integration for verifiable and scalable solutions.
PerPilot: Personalizing VLM-based Mobile Agents via Memory and Exploration
- PerPilot: introduces a plug-and-play LLM-powered framework for mobile agents, with Personalization Perception module (identifies personalized instructions, extracts elements), Personalization Completion module (retrieves/explores missing personalized information), Memory-based Retrieval (accesses stored user-specific information), Reasoning-based Exploration (infers apps, generates exploration instructions), and Agent Execution (executes clarified, explicit instructions), enabling autonomous perception, understanding, and execution of personalized user instructions.
- The framework leverages LLMs to identify personalized elements, first attempting to retrieve information from a Memory Database, and if unsuccessful, employing Reasoning-based Exploration to infer relevant apps and generate App Exploration Instructions to find missing data.
- PerPilot integrates with existing VLM-based mobile agent systems, progressively improving its personalization performance through continuous learning and memory updates, and is evaluated using the novel PerInstruct Dataset.
Neural Algorithmic Reasoners informed Large Language Model for Multi-Agent Path Finding
- LLM-NAR (Neural Algorithmic Reasoners informed Large Language Model): introduces a novel framework for Multi-Agent Path Finding (MAPF) that leverages neural algorithmic reasoners to enhance LLM's ability to process spatial map information, including an LLM for MAPF, a GNN-based NAR, and a cross-attention mechanism.
- The framework employs a tailored prompt interaction strategy for the LLM, a GNN-based NAR to capture map intricacies and spatial relationships, and a cross-attention mechanism to fuse LLM linguistic instructions with GNN spatial data.
- LLM-NAR significantly outperforms existing LLM-based approaches in solving MAPF problems by integrating GNNs with map information, demonstrating superior performance in both simulation and real-world experiments.
FinReflectKG: Agentic Construction and Evaluation of Financial Knowledge Graphs
- FinReflectKG (Reflection Driven Extraction Framework): introduces a robust and generalizable knowledge graph (KG) construction framework that integrates intelligent document parsing, table-aware semantic chunking, schema-guided iterative extraction, and a reflection-driven feedback loop to build a large-scale financial KG dataset from SEC 10-K filings.
- The framework supports three extraction modes—single-pass, multi-pass, and reflection-agent-based—with the latter achieving superior extraction quality through iterative refinement and a 64.8% compliance score.
- FinReflectKG also includes a comprehensive evaluation pipeline, combining rule-based checks, statistical validation, and LLM-as-a-Judge assessments to holistically measure extraction quality and advance financial KG research.
AgentRAN: An Agentic AI Architecture for Autonomous Control of Open 6G Networks
- AgentRAN (An Agentic AI Architecture for Autonomous Control of Open 6G Networks): introduces an AI-native, Open RAN-aligned agentic framework with AI Agents (LLM-powered autonomous entities), an AI-RAN Factory (Automated agent synthesis pipeline), a Data Lake (KPI and decision repository), an Agent-To-Agent (A2A) Protocol (Agent communication interface), a Model Context Protocol (MCP) (API discovery interface), a Context Repository (Aggregates agent information), dApps (Real-time RAN control logic), xApps (Near-real-time RAN adaptations), and rApps (Non-real-time RAN policies), enabling autonomous control of Open 6G networks through hierarchical intent decomposition and NL-based coordination.
- The framework's LLM-powered AI agents interpret natural language intents, negotiate strategies, and orchestrate control loops across various timescales, spatial domains, and protocol layers, replacing rigid APIs with flexible NL coordination.
- The AI-RAN Factory, leveraging the Data Lake, continuously generates and refines agents through code generation, model distillation, fine-tuning, and hybrid creation, transforming the network into a self-learning system that evolves its own intelligence.
RepoTransAgent: Multi-Agent LLM Framework for Repository-Aware Code Translation
- RepoTransAgent (Multi-Agent Large Language Model Framework): introduces a novel multi-agent LLM framework for repository-aware code translation, with RAG Agent (retrieves similar functions), Context Agent (gathers contextual information), and Refine Agent (translates, refines code iteratively), where it systematically decomposes the translation process into specialized subtasks.
- The framework leverages retrieval-augmented generation for contextual information, employs adaptive prompts tailored to varying repository scenarios, and integrates a reflection-based mechanism for systematic error correction.
- Evaluated on hundreds of Java-C# translation pairs, RepoTransAgent significantly outperforms state-of-the-art baselines in compile and pass rates, demonstrating robustness and generalizability across different LLMs.
Enhancing LLM-Based Social Bot via an Adversarial Learning Framework
- EvoBot (Evolving Large Language Model-based social Bot): introduces an LLM-based social bot enhanced through an adversarial learning framework, comprising EvoBot (generative LLM agent), an Adversarial Learning Framework (overall training paradigm), a Data Preparation Module (extracts/summarizes social data), a Supervised Fine-Tuning Module (initializes EvoBot), a Direct Preference Optimization Module (refines content), a Detector Module (co-adapting adversary), and an Evaluation Module (assesses performance).
- The framework initializes EvoBot via SFT on human social media data, then iteratively refines its human-like content generation using DPO, guided by feedback from a co-adapting Detector that concurrently improves its ability to distinguish bots from humans.
- This adversarial process creates an increasingly challenging learning environment for EvoBot, enabling it to generate content aligned with diverse user profiles, bypass detection, and accurately model real-world opinion dynamics and information spread in multi-agent simulations.
LLM-based Agentic Reasoning Frameworks: A Survey from Methods to Scenarios
- LLM-based Agentic Reasoning Frameworks Taxonomy: introduces a systematic taxonomy that decomposes agentic reasoning frameworks into single-agent, tool-based, and multi-agent methods, with all identifiable components and their roles.
- The survey provides a comprehensive review of key application scenarios, analyzes characteristic features of each framework, and summarizes different evaluation strategies.
- This work aims to offer a panoramic view to facilitate understanding of the strengths, suitable scenarios, and evaluation practices of diverse agentic reasoning frameworks.
Attacking LLMs and AI Agents: Advertisement Embedding Attacks Against Large Language Models
- AEA (Advertisement Embedding Attacks): introduces a new class of LLM security threats that stealthily inject promotional or malicious content into model outputs and AI agents, leveraging an Attacker (initiates malicious activity) to manipulate LLM Service Distribution Platforms (SDP) (distributes LLM inference) or Open-Source Model Distribution Platforms (MDP) (hosts open-source models) by injecting AEA Attack Data (malicious content) into the Attacked Backend Program (intercepts/modifies data) on a Computing Platform (executes LLM inference), ultimately affecting Users (receives tampered responses) and API Providers (provides LLM inference).
- The attack operates through two low-cost vectors: hijacking third-party service-distribution platforms to prepend adversarial prompts, or publishing back-doored open-source checkpoints fine-tuned with attacker data, causing models to return covert ads, propaganda, or hate speech.
- The paper also introduces a Prompt-Based Self-Inspection Defense Method (mitigates prompt attacks) to detect and defend against such attacks, highlighting an urgent gap in LLM security requiring coordinated responses.
- SonoCraftAR: introduces a proof-of-concept prototype empowering Deaf and hard-of-hearing (DHH) users to author personalized, sound-reactive AR interfaces by converting natural language User Prompts into animated Unity C# scripts via a multi-agent LLM pipeline (Prompt Enhancement, Code Generation, Code Checker agents), which are then compiled by Roslyn, rendered with the Shapes library, and dynamically animated by Real-time audio signal processing for display on HoloLens 2.
- The system extracts dominant frequency from continuous audio input using a Python server with FFT and NumPy, then maps this data to visual properties like size and color for dynamic AR interface animations.
- This approach demonstrates the feasibility of open-ended AR interface authoring for sound accessibility, allowing DHH users to create custom visualizations reflecting individual preferences.
TradingGroup: A Multi-Agent Trading System with Self-Reflection and Data-Synthesis
- TradingGroup: introduces a multi-agent trading system with a self-reflective architecture and an end-to-end data-synthesis pipeline, including News-Sentiment, Financial-Report, Stock-Forecasting, Style-Preference, and Trading-Decision Agents, a Risk-Management Module, a Self-Reflection Mechanism, a Data-Synthesis Pipeline, an LLM, Memory (Milvus), and Tools (Online Search), designed to address limitations in existing LLM-based trading systems.
- The system integrates performance metrics, agent logs, and risk signals into a coherent feedback loop for effective self-reflection and dynamic strategy optimization, enabling dynamic style switching and price forecasting.
- TradingGroup automatically collects and labels trading-process data to provide high-quality post-training samples for fine-tuning base LLMs, demonstrating superior performance over various baseline strategies in backtesting experiments.
- Neuro-Symbolic AI Framework: introduces a multi-agent, neuro-symbolic framework designed for educational support, featuring an Educational Ontology, a Tutor Agent, and a Peer Agent, interacting within Digital Learning Environments with Students.
- This framework addresses generalizability, educational effectiveness, and the social learning gap by unifying specialized agents under a coherent architecture, enabling cross-domain applicability and grounding LLM dialogue.
- The system leverages a symbolic knowledge base (Educational Ontology) for verifiable structure and neural agents (Tutor and Peer) for adaptive, generative power, ensuring scalable and pedagogically sound interactions.
- DCCM (Data-Centric Criticality Metrics): introduces a data-centric approach for robust offline Reinforcement Learning in autonomous motion planning by augmenting Conservative Q-Learning (CQL) with a Data Curation Pipeline that employs Criticality Metrics (Heuristic-Based, Uncertainty-Based, Behavior-Based) and non-uniform Data Sampling Mechanisms to train a Goal-Conditioned, Shared-Encoder Actor-Critic Architecture.
- The framework addresses the long-tail problem in real-world driving logs by focusing the learning process on information-rich samples, significantly reducing safety-critical failures like collisions and off-road incidents compared to uniform data sampling.
- Data-driven criticality metrics, particularly those based on model uncertainty and expert action rarity, demonstrate superior performance in improving core safety and goal achievement over human-defined heuristics, with timestep-level weighting excelling in reactive safety and scenario-level in long-horizon planning.
Experiences with Model Context Protocol Servers for Science and High Performance Computing
- MCP (Model Context Protocol): introduces an architecture for AI agents to discover, invoke, and coordinate scientific capabilities across heterogeneous cyberinfrastructure, leveraging LLMs for planning and execution.
- The architecture integrates various MCP servers for services like data transfer, compute, search, facility status, event streaming, and machine learning/bioinformatics tools, enabling agents to orchestrate complex, multi-site scientific workflows.
- The approach emphasizes building thin MCP adapters over existing services, separating discovery from invocation, and allowing agents to dynamically generate glue code, enhancing resilience and recovery for long-running tasks.
The AI in the Mirror: LLM Self-Recognition in an Iterated Public Goods Game
- Iterated Public Goods Game Simulation: introduces a study analyzing LLM self-recognition and cooperation, with LLM Agents (game players) interacting in a Game Environment (iterated public goods game) guided by System Prompts (agent behavior directives) over Game Rounds (repeated interaction cycles), using a Contribution Mechanism (agent point allocation) and Payoff Calculation (individual reward determination), supported by a Multiplier (common pool amplification), Context Window (agent historical memory), and for Study 1, a Sentiment Analysis Module (reasoning text scorer) and Spearman Correlation Module (statistical relationship analyzer).
- The simulation investigates how LLMs behave under "no-name" (playing against "another AI agent") versus "name" (playing against "themselves") conditions, and with "neutral," "collective," or "selfish" objectives, measuring point contributions as a proxy for cooperation or defection.
- Findings indicate that informing LLMs they are playing against themselves significantly alters their cooperation tendencies, with more defection under "collective" prompts and more cooperation under "selfish" prompts in the "name" condition, highlighting the influence of perceived identity on AI agent behavior.
LLM-Driven Intrinsic Motivation for Sparse Reward Reinforcement Learning
- LLM+VAE strategy: introduces a novel approach for sparse reward reinforcement learning, combining a Variational AutoEncoder (VAE) for state novelty-based intrinsic rewards and an LLM for goal-oriented intrinsic rewards, which are then aggregated with extrinsic rewards to guide an Actor-Critic (A2C) agent.
- This combined strategy addresses sparse reward challenges by leveraging VAE for exploration of new states and LLM's pre-trained knowledge to facilitate progressive exploitation towards goals.
- The framework computes a total reward signal from extrinsic, VAE-derived, and LLM-derived intrinsic rewards, enabling the A2C agent to learn effectively in environments where traditional methods fail.
TRAINING LANGUAGE MODEL AGENTS TO FIND VULNERABILITIES WITH CTF-DOJO
- CTF-FORGE (Automated Pipeline for CTF Challenge Environment Creation): introduces an automated pipeline for transforming publicly available CTF artifacts into ready-to-use execution environments, with Source (input artifacts for challenges), Rehost (LLM input for environment generation), Language Model (generates configuration files), Heuristic Rules (guides LLM generation), Dockerfile (builds runtime, embeds flags), Docker Compose (configures Docker services/networks), Challenge JSON (describes challenge structure, flag verification), CTF Challenge Runtime (containerized execution environment), and Cybersecurity Agent (interacts with runtime to solve challenges).
- This pipeline leverages LLMs to automatically generate Docker-based runtime environments for CTF-DOJO, enabling scalable and reproducible training of cybersecurity agents.
- CTF-FORGE significantly reduces the manual effort and time traditionally required for setting up CTF challenges, achieving a high success rate in creating stable and executable environments.
- Interactive Graph Visualization Framework: introduces an interactive system for the CM4AI KG, integrating WebGL visualization with LLM agents to enable responsive exploration, filtering, and AI-driven recommendations with justifications for large scholarly knowledge graphs.
- The system leverages Specter2 for author and dataset embeddings, t-SNE and UMAP for dimensionality reduction, and PixiJS for large-scale interactive node visualization, overcoming limitations of traditional graph tools.
- It features a multi-agent LLM-powered CM4AI MATRIX for expertise-gap based teaming recommendations, including an expertise gap detection agent and a reranking agent, to identify potential collaborators and dataset users.
- Interactive Graph Visualization Framework: introduces an interactive system for the CM4AI KG, integrating WebGL visualization with LLM agents to enable responsive exploration, filtering, and AI-driven recommendations with justifications for large scholarly knowledge graphs.
- The system leverages Specter2 for author and dataset embeddings, t-SNE and UMAP for dimensionality reduction, and PixiJS for large-scale interactive node visualization, overcoming limitations of traditional graph tools.
- It features a multi-agent LLM-powered CM4AI MATRIX for expertise-gap based teaming recommendations, including an expertise gap detection agent and a reranking agent, to identify potential collaborators and dataset users.
- Interactive Graph Visualization Framework: introduces an interactive system for the CM4AI KG, integrating WebGL visualization with LLM agents to enable responsive exploration, filtering, and AI-driven recommendations with justifications for large scholarly knowledge graphs.
- The system leverages Specter2 for author and dataset embeddings, t-SNE and UMAP for dimensionality reduction, and PixiJS for large-scale interactive node visualization, overcoming limitations of traditional graph tools.
- It features a multi-agent LLM-powered CM4AI MATRIX for expertise-gap based teaming recommendations, including an expertise gap detection agent and a reranking agent, to identify potential collaborators and dataset users.
How to cite my work?
@misc{MaattaAutonomousAgents2023,
author = {Teemu Maatta},
title = {Autonomous Agents},
year = {2023},
howpublished = {\url{http://github.com/tmgthb/Autonomous-Agents}},
note = {Accessed: YYYY-MM-DD}
}