JEPA-Anything: One World Model Framework Spanning Molecules, Cells, Robots and Weather
burny_tech · x · 2026-10-06
Researchers introduced JEPA-Anything, a domain-agnostic world modeling framework built on orthogonal predictive factorization (OPF), an extension of joint-embedding predictive architectures that decomposes latent targets into complementary factors learned via dedicated pathways and recombined in a shared predictive design.
It was evaluated across seven domains — vision, biology, clinical trajectories, control, molecular dynamics, physical fields, and weather:
- Beats matched JEPA baselines on all 10 matched dynamics tasks
- Cuts single-intervention prediction error on Interventional Pong by 34.8%
- Forecasts over 1,000 clinical events
- Achieves the lowest one-step and 100-step molecular rollout errors in all four tested systems
Paper and code are available; the authors argue a single learning principle can support world modeling across radically different systems.
More from Research
- CC-Bench at COLM 2026 finds LLMs still default to stereotypes over implicit cultural cues — MaartenSap · 2026-10-06
- UCLA PhD's LLM agents produce 126K-line Lean 4 proof of MIP* = RE core theorem in 63 days — siyan_zhao · 2026-10-06
- UniReps 4th edition heads to Paris on Dec 12, 2026, with submissions extended to Oct 10 — ClementineDomi6 · 2026-10-06
- AI-generated key-value stores beat general-purpose DBs via specialized architecture — CShorten30 · 2026-10-06
- Real2sim's last mile for robotics: verification pipeline for physics and affordance — chris_j_paxton · 2026-10-06
- LenVM lands COLM 2026 Spotlight: value model predicts remaining generation length for efficient reasoning — xwang_lk · 2026-10-06