1 paper
Arman Akbari, Ci Zhang, Arash Akbari +6
World-action models (WAMs) jointly generate future video and robot actions through iterative diffusion, achieving strong performance on manipulation benchmarks but requiring tens o…