Oxford x NUS Introduce Mental World Modeling to Boost Human Decision Prediction
jiqizhixin · x · 2026-08-22
Researchers from Oxford and NUS introduce Mental World Modeling (MWM), addressing the blind spot where current AIs track physical motions but fail to understand underlying intentions.
The Problem: Existing world models cannot distinguish motivations behind actions (e.g., reaching for a knife to cook vs. attack).
The Solution: MWM forces world models to track hidden beliefs, desires, and social rules alongside physical scenes. The baseline model, MENTIS, decomposes this into 5 steps: parsing state, rendering target-specific views, splitting actions, updating physical and mental states synchronously, and scoring branches.
Results: Tested across text, image, and video scenarios with 8 modern LLM world models, explicit mental-state tracking consistently outperforms physical-only baselines in predicting human decisions. The bottleneck remains that current models struggle to infer intentions from messy sensory input.
More from Research
- WEDGE paper generates training environments conditioned on navigation agent failures — _akhaliq · 2026-10-16
- RExBench: best coding agent completes only 33% of autonomous research extensions — najoungkim · 2026-10-16
- Google's CLASHarness benchmarks multi-agent conflicts in computer use, finds frontier VLMs struggle — google · 2026-10-16
- Tsinghua team proves vanilla flow-matching sampling can't exceed constant-depth expressive power — SQZ · 2026-10-16
- APEX-Harm benchmark: script-based injection attacks easily compromise agents in shared workspaces — KAUST · 2026-10-16
- Residual-stream forecasts predict agent violations before they happen, USC study shows — USC · 2026-10-16