joint-embedding

3 posts tagged with "joint-embedding"

July 9, 2026

V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning

Internet-scale JEPA video model pretrained on 1M+ hours of video, achieving SOTA understanding benchmarks and enabling zero-shot robotic planning via action-conditioned post-training with <62h robot data

world-models representation-learning robotics joint-embedding mask-denoising

July 9, 2026

V-JEPA 2.1: Unlocking Dense Features in Video Self-Supervised Learning

Next-gen JEPA model with dense predictive loss, deep self-supervision, and multi-modal tokenizers achieving SOTA on dense visual tasks (depth, segmentation, STA) while retaining global scene understanding

representation-learning video-understanding robotics joint-embedding mask-denoising

July 9, 2026

LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels

First stable end-to-end JEPA that trains from raw pixels using only two loss terms (prediction + SIGReg), with ~15M params on single GPU, 48x faster planning than foundation-model world models

world-models representation-learning reinforcement-learning joint-embedding anti-collapse