MODA: MOdular Duplex Attention for Multimodal Perception, Cognition, and Emotion Understanding

Zhicheng Zhang^1,2,†, Wuyou Xia¹, Chenxi Zhao^1,†, Yan Zhou³, Xiaoqiang Liu³, Yongjie Zhu^3,‡, Wenyu Qin³, Pengfei Wan³, Di Zhang³, Jufeng Yang^1,2,✉

¹Nankai University      ²Pengcheng Laboratory      ³Kuaishou Technology     

^†Work done at KlingAI      ^‡Project Leader      ^✉Corresponding Author     

🎉 Accepted by ICML 2025 Spotlight 🎉

[📃 Paper ] [📦 Code ] [⚒️ Project ] [📅 Slide ]

TL;DR: We i) identify attention deficit disorder as a critical barrier hindering fine-grained content understanding in MLLMs; ii) introduce a modular duplex attention mechanism to mitigate modality bias and enhance attention score justification; and iii) develop MODA-based MLLMs that enable fine-grained multimodal understanding across perception, cognition, and emotion tasks.

📈 1. News

• 🔥2025-07-10: Creating repository. The code is uploading ...
• 2025-05-01: MODA has been accepted to ICML 2025！