FactorJEPA: A New World Model for Crowded Urban Environments
Kapil Wanaskar · hf · 2026-08-08
Existing world models are typically evaluated on low-density, lane-structured settings and struggle with populous, chaotic urban environments. To bridge this gap, researchers introduced the DENSEWORLD regime and released a large-scale dataset comprising 1,000 hours of drive-through, walk-through, and aerial video across 22 cities.
The paper also proposes FactorJEPA. Instead of encoding the future in a monolithic latent space, this architecture decomposes scenes into layout, entities, and interactions using separated subspaces and a visibility gate to preserve partially observed agents. Experiments demonstrate improvements in predictive accuracy, causal sensitivity, and robustness, with method rankings replicating consistently across 2B and 1B V-JEPA 2.1 backbones.
More from Research
- AI will be superhuman at science regardless, and Europe can't opt out — teortaxesTex · 2026-10-16
- SERO augments long-horizon agentic RL with executable oracle continuations — Shanghai-AI-Laboratory · 2026-10-16
- Tsinghua and Amap propose OmniVerifier-C1, a graded verifier beating binary judges — Bowei Liu · 2026-10-16
- Alibaba open-sources MAI-Gym, an all-in-one GUI agent eval and RL training platform — Tongyi-MAI · 2026-10-16
- Peking University's SimForcing distills simulation motion priors into robot world models — PekingUniversity · 2026-10-16
- CoWorld Reasoner uses an evolving video world model as a visual reasoning engine — ModalityDance · 2026-10-16