Native Action-Prior Learning from Videos for World Action Models Paper • 2610.03391 • Published Oct 2 • 56
GenFirst: Generation Before Reconstruction for Stable End-to-End Latent Generative Modeling Paper • 2608.29335 • Published Aug 29 • 74
From Activation to Causality: Discovery of Causal Visual Representations in the Human Brain Paper • 2605.23895 • Published May 22 • 53
BRIDGE: Background Routing and Isolated Discrete Gating for Coarse-Mask Local Editing Paper • 2605.07846 • Published May 11 • 1