JEPA-Anything: Learning Predictive Models across Different Worlds
Researchers present JEPA-Anything, a domain-agnostic world-modeling framework improving prediction across vision, biology, control, molecular dynamics, and weather.
JEPA-Anything extends joint-embedding predictive architectures with orthogonal predictive factorization, decomposing latent targets into complementary factors learned through dedicated pathways. It was evaluated across seven domains, including 10 matched dynamics tasks, forecasting of over 1,000 clinical events, and 100-step molecular rollouts across four systems. Against matched JEPA baselines it improves metrics on all 10 dynamics tasks and cuts single-intervention prediction error on Interventional Pong by 34.8%. A factor-nominated biological intervention received experimental support in cell co-cultures, patient-derived organoids, tumor fragments, and mice.
- Orthogonal predictive factorization decomposes latent targets into complementary factors
- Evaluated across seven domains from vision to weather
- Improves all 10 dynamics tasks and cuts Interventional Pong error by 34.8%
- Biological intervention validated in organoids, tumor fragments, and mice
- Products
- JEPA-Anything
- Organizations
- Gen-Verse
From the original article
1 min read at huggingface.co
World modeling enables intelligence to anticipate consequences, guide interventions, and learn from interaction. Yet predictive models remain domain-specific: can a common learning principle support world modeling across radically different systems? We introduce JEPA-Anything, a domain-agnostic framework based on orthogonal predictive factorization (OPF).
By Taoyong Cui, Zhongyao Wang, Xinyue Xu · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.