JEPA-Anything: Learning Predictive Models across Different Worlds

World modeling gives intelligence the ability to anticipate consequences, guide interventions, and learn from interaction, but predictive models today remain domain-specific. The authors ask whether a single common learning principle can support world modeling across radically different systems, and introduce JEPA-Anything to test that idea.

The method is orthogonal predictive factorization (OPF), a domain-agnostic framework that extends joint-embedding predictive architectures. OPF decomposes latent targets into complementary factors, learns each through dedicated pathways, and recombines them within a shared predictive design.

Evaluation spans seven domains — vision, biology, clinical trajectories, control, molecular dynamics, physical fields, and weather — and covers representation learning, intervention prediction, out-of-distribution generalization, and long-horizon dynamics. The experimental suite includes 10 matched dynamics tasks, forecasting of over 1,000 clinical events, and 100-step molecular rollouts across four systems.

Against matched JEPA baselines, JEPA-Anything improves reported metrics on all 10 dynamics tasks and cuts single-intervention prediction error on Interventional Pong by 34.8%. It also attains the lowest one-step and 100-step molecular errors among compared methods in all four systems. Beyond prediction, a factor-nominated biological intervention receives experimental support in cell co-cultures, patient-derived organoids, tumor fragments, and mice, and latent orbital modes recover the Keplerian scaling exponent with a fitted slope of -1.4991.

Together the results support a common factorized predictive principle across heterogeneous worlds, linking world modeling to intervention and to experimentally grounded scientific discovery. Code is available at https://github.com/Gen-Verse/JEPA-Anything.