I'm Rong Ding, a researcher and model developer with a keen interest in how language and meaning are processed by humans and by biologically inspired systems/models (e.g., AI). I build computational models that bridge natural language, human data (behavioural/neural), and structured semantics.

I believe in interpretable AI, interdisciplinary thinking/knowledge integration, and building tools that help people reason better. When not working, I enjoy reading about complex systems—the kind that tries to explain neural networks, language, and also black holes and traffic jams. I also search for the meaning of meaning during hiking, ballet dancing and travelling.

• How symbolic representations (ontologies, knowledge graphs) can be combined with semantic embeddings and neural networks
• How AI models augment reconstruction and symbolic representation of non-linear system dynamics (e.g., in the brain)

• sql-explore-semantica: SQL-driven semantic modeling for user funnel analysis with BigQuery. Designed to demonstrate practical skills in data modelling, exploratory analysis, and semantic abstraction, aligning with Analytics Engineering principles.
• Semantics_meets_Ontology: bridging distributional semantics and biological ontologies to explore whether embedding models can recover and reason about relationships within and between genes (GO) and plant traits (PTO). On the way to enable scalable, zero-shot categorisation of novel traits and to facilitate ontology-aware gene–phenotype predictions, contributing to both plant biology and semantic AI applications
• ModellingHumanReading: Comparing Word2Vec models in handling word semantics, and SRN vs LSTM models in characterising the statistical structure of language and syntactic ambiguity (garden-path sentences). Providing significant insights into human neuro-cognitive behaviours and helping improve interactive language learning systems
• DecodingUnderPartialObservability: Building Autoencoders that effectively reconstruct latent states and system dynamics under varying degrees of (partial) observability. Offering valuable insights for neural signal analysis and clinical interpretation, and also improving the design of brain–computer interfaces (BCIs) by enhancing the temporal decoding of neural activity
• Coupled_STiMCON: A neural (oscillatory) model that combines internal language knowledge (predictability) with external inputs in speech processing. Advancing understanding of language processing, inspiring more adaptive mechanisms in AI, and informing neuromorphic designs for predictive, human-like learning in technologies

• PhD Language and Computation in Neural Systems @ Max Planck Institute for Psycholinguistics & Donders Institute, Radboud University (2020-2025)
• MSc Language Sciences (with specialisation in neuroscience and communication) @ University College London (2018-2019)
• BSc Psychology & BA Indonesian Language and Literature @ Peking University (2014-2018)

• E-mail: sariahding1007@gmail.com