MLLM as Retriever: Interactively Learning Multimodal Retrieval for Embodied Agents
机构 * School of Computer Science, Peking University(北京大学计算机科学学院) ; Institute of Software, Chinese Academy of Sciences(中国科学院软件研究所) ; Beijing Academy of Artificial Intelligence(北京人工智能研究院)
专题命中 机器人操作 :embodied agent(title,abstract);分类 cs.LG
Comments ICLR 2025