Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization
- Published
- Source
- arXiv
- Paper number
- 1079
- Field
- Machine Learning
- arXiv ID
- 2609.10464
Key points
- SG-JEPA takes a gravity condition and recursively predicts multi-step latent states, jointly training the encoder and predictor.
- Generalization is tested on eight MuJoCo datasets that train in a narrow gravity range and evaluate over a wider one.
- The paper reports that 2D prediction error falls to as little as half that of DINO-WM and 3D robot-control success rates rise by up to 2.5 times.
- Training a new predictor on a frozen encoder shows the key gain lies in dynamics-relevant representations, which are then used as control inputs for a separate diffusion policy.
- Settings that vary several physical variables beyond gravity together were not validated, and transfer under changed object shape and rotation is uneven.
Paper links
External research summaries. These are not HDATF publications or measured product results.