1 paper
Jiarui Yang, Wen Huang, Jiale Zhang +2
Vision-Language-Action (VLA) models have become the dominant recipe for generalist manipulation, yet they are almost universally trained by behavior cloning: a policy imitates expe…