AI RESEARCH
LaST-R1: Reinforcing Action via Adaptive Physical Latent Reasoning for VLA Models
arXiv CS.CV
•
ArXi:2604.28192v1 Announce Type: cross Vision-Language-Action (VLA) models have increasingly incorporated reasoning mechanisms for complex robotic manipulation. However, existing approaches share a critical limitation: whether employing explicit linguistic reasoning that suffers from latency and discretization, or utilizing expressive continuous latent reasoning, they are predominantly confined to static imitation learning that limits adaptability and generalization. While online reinforcement learning (RL) has been