VLA-JEPA. [ECCV 2026] VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Model
world-model-VLA-VLN.
