Jump-Start Reinforcement Learning with Vision-Language-Action Regularization

ArXi:2604.13733v1 Announce Type: cross Reinforcement learning (RL) enables high-frequency, closed-loop control for robotic manipulation, but scaling to long-horizon tasks with sparse or imperfect rewards remains difficult due to inefficient exploration and poor credit assignment. Vision-Language-Action (VLA) models leverage large-scale multimodal pre