ACL Findings2026
MemWeaver: Weaving Hybrid Memories for Traceable Long-Horizon Agentic Reasoning
Juexiang Ye, Xue Li, Xinyu Yang, Chengkai Huang, Lanshun Nie, Lina Yao, Dechen Zhan
Abstract
Large language model-based agents operating in long-horizon interactions require memory systems that support temporal consistency, multi-hop reasoning, and evidence-grounded reuse across sessions. Existing approaches largely rely on unstructured retrieval or coarse abstractions, which often lead to temporal conflicts, brittle reasoning, and limited traceability. We propose MemWeaver, a unified memory framework that consolidates long-term agent experiences into three interconnected components: a temporally grounded graph memory for structured relational reasoning, an experience memory that abstracts recurring interaction patterns from repeated observations, and a passage memory that preserves original textual evidence. MemWeaver employs a dual-channel retrieval strategy that jointly retrieves structured knowledge and supporting evidence to construct compact yet informationdense contexts for reasoning. Experiments on the LoCoMo benchmark demonstrate that MemWeaver substantially improves multi-hop and temporal reasoning accuracy while reducing input context length by over 95% compared to long-context baselines. Our data and code are available at here.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
- Improving Language Models by Retrieving from Trillions of Tokens cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
- A-Mem: Agentic Memory for LLM Agents cites Improving Language Models by Retrieving from Trillions of Tokens
- A-Mem: Agentic Memory for LLM Agents cites Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
Builds on14
- Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksPatrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni et al.NeurIPS 2020 · 19,162 citations
- Generative Agents: Interactive Simulacra of Human BehaviorJoon Sung Park, Joseph C. O'Brien, Carrie Jun Cai, Meredith Ringel Morris et al.UIST 2023 · 1,882 citations
- Improving Language Models by Retrieving from Trillions of TokensSebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai et al.ICML 2022 · 1,629 citations
- A-Mem: Agentic Memory for LLM AgentsWujiang Xu, Zujie Liang, Kai Mei, Hang Gao et al.NeurIPS 2025 · 1,138 citations
- HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language ModelsBernal Jimenez Gutierrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga et al.NeurIPS 2024 · 395 citations
Related papers
- GAM: Hierarchical Graph-based Agentic Memory for LLM AgentsZhaofen Wu, Hanrong Zhang, Fulin Lin, Wujiang Xu et al.ACL 2026 · 8 citations
- Lightweight LLM Agent Memory with Small Language ModelsJiaquan Zhang, Chaoning Zhang, Shuxu Chen, Zhenzhen Huang et al.ACL 2026 · 4 citations
- LiCoMemory: Lightweight and Cognitive Agentic Memory for Efficient Long-Term ReasoningZhengjun Huang, Zhoujin Tian, Qintian Guo, Fangyuan Zhang et al.ACL Findings 2026 · 2 citations
- Chain-of-Memory: Lightweight Memory Construction with Dynamic Evolution for LLM AgentsXiucheng Xu, Bingbing Xu, Tian Xueyun, Zihe Huang et al.ACL 2026 · 4 citations
- Memory Matters More: Event-Centric Memory as a Logic Map for Agent Searching and ReasoningYuyang Hu, Jiongnan Liu, Jiejun Tan, Yutao Zhu et al.ACL Findings 2026
