SOURCE-LINKED INTELLIGENCE

MotionJEPA: Preventing Temporal Feature Collapse by Capturing Visual Changes in Latent Space

arXiv · AI, language, vision and robotics · article · Sep 20, 2026 · UTC

Joint Embedding Predictive Architectures (JEPAs) are a promising paradigm for learning task-agnostic latent world models without visual reconstruction. However, standard JEPA training exhibits a strong inductive bias towards slow features, causing feature suppression and the collapse of latent representation. While inverse dynamics provides temporal anti-collapse, it relies on action labels and offers little incentive to embed general, unlabeled dynamics. We introduce Difference Image and Single image embedding Regularization (DISReg), a novel regularizer that builds on an inverse-dynamics-sty

Read original source ↗ Open in workspace

recordType

paper

region

Global