MemBodied: Recurrent Associative Memory for Vision-Language-

arXiv:2609.28256v1 Announce Type: cross Abstract: Vision-Language-Action models provide a strong foundation for general-

推荐理由：研究进展具备参考价值，值得关注