AI RESEARCH

Entity-Centric World Models: Interaction-Aware Masking for Causal Video Prediction

arXiv CS.CV

ArXi:2605.15466v1 Announce Type: new Learning predictive world models from unlabelled video is a foundational challenge in artificial intelligence. While Joint Embedding Predictive Architectures (JEPA) have set new benchmarks in semantic classification, they often remain physics-blind, failing to capture the causal dynamics necessary for downstream reasoning. We hypothesize that this stems from standard patch-based masking strategies, which prioritize visual texture over rare but informative kinematic events.