Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Group Cognition Learning: Making Everything Better Through Controlled Two-Stage Agents Collaboration
Chunlei Meng, Pengbin Feng, Rong Fu, Hoi Leong Lee, Xiaojing Du, Zhaolu Kang, Zeyu Zhang, Weilin Zhou, Chun Ouyang, Zhongxue Gan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add selective interaction gating instead of full exchange (ablation) | GCL · CMU-MOSI | MAE: 0.721 -> 0.685 | Improved |
| add Routing Agent (ablation) | GCL · CMU-MOSI | MAE: 0.694 -> 0.685 | Improved |
| add Auditing Agent (ablation) | GCL · CMU-MOSI | MAE: 0.699 -> 0.685 | Improved |
| add Public-Factor Agent (ablation) | GCL · CMU-MOSI | MAE: 0.702 -> 0.685 | Improved |
| use contribution-aware aggregation weights instead of uniform weights (ablation) | GCL · CMU-MOSI | MAE: 0.698 -> 0.685 | Improved |
| add gain alignment loss L_gain (ablation) | GCL · CMU-MOSI | MAE: 0.698 -> 0.685 | Improved |
| add redundancy control loss L_red (ablation) | GCL · CMU-MOSI | MAE: 0.703 -> 0.685 | Improved |
| add public factor loss L_pub (ablation) | GCL · CMU-MOSI | MAE: 0.696 -> 0.685 | Improved |
| add governance losses (gain, redundancy, public) to the task loss (ablation) | GCL · CMU-MOSI | MAE: 0.712 -> 0.685 | Improved |
| add acoustic and visual modalities to language (ablation) | GCL · CMU-MOSI | MAE: 0.714 (only Language) -> 0.685 | Improved |
| include language modality (versus dropping it) (ablation) | GCL · CMU-MOSI | MAE: 0.905 -> 0.685 | Improved |
| add redundancy control under message permutation stress (ablation) | GCL · message permutation stress test | task accuracy and coupling (HSIC/CKA) | Improved |
Closest papers by meaning
- ModularAgent: A Task-Aware Modular Framework for Joint Optimization of Multimodal Large Language Models and World Models (CVPR 2026 · similarity 0.86)
- Enhancing Multimodal Cooperation via Sample-level Modality Valuation (CVPR 2024 · similarity 0.85)
- Test-Time Matching: Unlocking Compositional Reasoning in Multimodal Models (ICLR 2026 · similarity 0.85)
- NegoCollab: A Common Representation Negotiation Approach for Heterogeneous Collaborative Perception (NeurIPS 2025 · similarity 0.85)
- R-C2: Cycle-Consistent Reinforcement Learning Improves Multimodal Reasoning (CVPR 2026 · similarity 0.85)
- SynIB: Informational Bottleneck for Maximizing Synergy in Multimodal Learning (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
