UCSD NeurIPS paper: modeling partner mental states boosts multi-agent performance
CatAstro_Piyush · x · 2026-10-18
A UCSD team (Hanan Gani, Lulu Shao, Manmohan Chandraker) released their NeurIPS paper Mental Models for Multi-Agent Systems.
- Problem: multi-agent interaction is partially observable — dialogue reveals actions but not the other agent's goals, knowledge, constraints, or expectations, so policies trained on surface behavior can pick plausible-but-strategically-wrong responses.
- Method: infer a compact latent belief state from interaction history, with explicit subspaces for first- and second-order beliefs, intent, and thoughts; jointly train a mental-model-conditioned utility so partner representations directly help evaluate candidate actions; the final policy runs standalone at inference.
- Results: explicit mental-state modeling improves downstream multi-agent performance on both language-only and multimodal benchmarks, spanning negotiation, cooperation, competition, and assistance.
Paper and code are available.
More from Research
- Laude Institute launches FrontierCS 2: agents must beat deleted paper contributions blind — lateinteraction · 2026-10-23
- RFD4-Proteina's 'underrated' docs detail partial diffusion for protein redesign — AllThingsApx · 2026-10-23
- MPI-Sec PhD Thesis on Vulnerability Detection Benchmarks Includes Pre-LLM Classic "Top Score on the Wrong Exam" — mboehme_ · 2026-10-23
- LiveNerf's 29-day data finds no nerfing of Claude Opus 5.5 — TheOnlyVibemaster · 2026-10-23
- New paper makes 9 falsifiable predictions toward a major AI-enabled terror attack by 2027-2028 — StephenLCasper · 2026-10-23
- HumanoidPose recovers humanoid robots' full-body joint states from RGB video alone — cholz · 2026-10-23