具身开发者社区

外观

论文HF Daily Papers · arXiv 2610.307556 天前

Raw-RAG在对话记忆评测中超越复杂智能体记忆系统 ​

研究显示基于原始历史的朴素RAG在LoCoMo等四大基准测试中六项排名第一，超最强智能体记忆0.13-5.53分。证实高证据覆盖率可弥补利用率不足，挑战了复杂记忆架构在对话任务中的必要性。

AI 整理
  • 数据与采集
  • 具身模型

本文由 AI 根据公开内容整理，可能有误，请以原文为准：Is Complex Agentic Memory Always Needed? Naive RAG Is Stronger Than We Thought（HF Daily Papers · arXiv 2610.30755）

核心内容 ​

针对LLM长期对话理解中复杂智能体记忆系统的必要性，研究提出Raw-RAG基线，仅对原始历史进行密集检索与重排序。在LoCoMo、LongMemEval、MEME及MemoryAgentBench四大基准的七种模型设置中，该方案六次位列第一，相比最强对比系统领先0.13至5.53分，确立了朴素RAG作为关键基线的地位。

机制分析表明，Raw-RAG凭借对必要证据（含答案值及推导依赖路径）的高覆盖率，有效抵消了信息利用率的劣势。实验发现即便完全访问经变换的结构化记忆存储，准确率仍低于原始证据检索。这揭示了在特定对话任务中，保留原始信息完整性比引入变异或结构化组织等复杂操作更具效能优势。

关键看点 ​

  • 评测表现：Raw-RAG在四大对话记忆基准的七种设置中六次夺冠，较最强智能体记忆系统高出0.13-5.53分
  • 核心机制：原始历史检索通过覆盖答案值及推理依赖路径等高价值证据，以高覆盖率弥补了低利用率缺陷
  • 工程启示：完整访问变换后的结构化记忆仍存在精度差距，验证了对话场景中保留原始信息优于复杂记忆重构
  • 基线价值：挑战了复杂代理记忆的普适必要性，确立基于密集检索与重排序的朴素RAG为具身对话必备基线

工程与产业观察 ​

该研究为具身智能长程交互提供重要工程参照：在构建机器人对话与任务规划系统时，应优先验证基于原始多模态数据流的检索增强基线。盲目引入复杂记忆重构可能引入信息损耗，保留原始观测数据的完整性往往是提升长尾任务鲁棒性的更优解，有助于降低端侧部署算力开销。

英文摘要 (Abstract) ​

LLM memory is widely used in long-term conversation understanding. Agentic memory systems often use mutation, structural organization, learned retrieval strategies, or other complex operations. Is this complexity always necessary? We study Raw-RAG, a naive RAG baseline using dense retrieval and reranking over raw history to supply evidence to a reader LLM. Across LoCoMo, LongMemEval, MEME, and MemoryAgentBench, Raw-RAG ranks first among the compared memory systems in six of seven benchmark-model settings, outperforming the strongest compared agentic memory system by 0.13-5.53 points. We investigate why raw-history retrieval remains competitive. Evidence analysis shows that its higher coverage of necessary evidence can offset lower utilization. Useful evidence includes both answer values and the relations and dependency paths needed to derive them. Even complete access to transformed memory stores leaves an accuracy gap relative to raw evidence, highlighting the importance of preserving raw information. These findings challenge the necessity of complex agentic memory for the evaluated conversation tasks and establish naive RAG as an essential baseline.

Hugging Face Daily Papers 热度：21 upvotes 作者：Zhengye Bao, shengtao zhang, Haokai Zhang, Shiqi Lou, Muning Wen, Minnan Luo

来源

本页事实对应的公开页面，以原文为准。

  1. HF Daily Papers · arXiv 2610.30755：Is Complex Agentic Memory Always Needed? Naive RAG Is Stronger Than We Thought访问 2026-10-25