#VAE (2)

FM-VLA: Force-based Memory for Vision-Language-Action Models in Contact-Rich Manipulation

VLA力觉感知触觉记忆增强接触丰富操作

Robust-WAM: Bridging Generative Pretraining and Semantic Foresight in World-Action Models

世界模型世界-动作模型视频生成VAE语义前瞻