Task-Adaptive Retrieval over Agentic Multi-Modal Web Histories via Learned Graph Memory
基于学习图记忆的代理多模态网络历史任务自适应检索
机构 * School of Engineering, Royal Melbourne Institute of Technology University(皇家墨尔本理工大学工程学院)
专题命中 跨模态检索 :multi-modal(title,abstract);分类 cs.AI
AI总结 本文提出ACGM方法,通过任务自适应图记忆检索代理历史,利用策略梯度优化提升检索质量,在多个数据集上取得显著成果。
Comments The 49th International ACM SIGIR Conference on Research and Development in Information Retrieval