1 paper
Leesai Park, Jiho HOng, Sanghyun Kim
Diffusion policies generate multimodal robot action sequences from demonstrations, but steering them toward deployment-time constraints typically relies on differentiable guidance…