Rolling-WAM: World Action Models with Rolling Imagination Paper • 2609.30247 • Published 7 days ago • 4
FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders Paper • 2609.31620 • Published 6 days ago • 134
Learning from Massive Human Videos for Universal Humanoid Pose Control Paper • 2412.14172 • Published Dec 18, 2024 • 10