Elamur Research⌕
☰
home / papers / 2026 / MemWeaver: Weaving Hybrid Memories for Traceable Long-Horizon Agentic Reasoning

MemWeaver: Weaving Hybrid Memories for Traceable Long-Horizon Agentic Reasoning

Paper · 2026 · Juexiang Ye, Xue Li, Yang Xinyu et al.

Summary

Conversational LLM agents that run across many sessions need memory that keeps track of when things happened, supports chaining several facts together, and lets answers be checked against what was actually said. Current memory designs fall short. Flat vector retrieval has no relational or temporal structure. Knowledge-graph memories accumulate noisy and conflicting extractions. Summaries of past behaviour are hard to verify. MemWeaver addresses this by consolidating dialogue history into three linked layers:

  • a knowledge graph with normalized absolute timestamps and session-level reconciliation;
  • experience items induced only from coherent clusters of repeated interactions;
  • a passage store keeping the raw text, attached to graph entities so it can be traced.

At answer time, two retrieval channels work together. One pulls graph triples, and the other gathers linked passages, experience items and a few globally retrieved passages. Their outputs are merged into a short context. On the LoCoMo benchmark, the system improves multi-hop and temporal question answering while using far less input than long-context baselines. This makes it attractive for long-running or resource-limited deployments.

Key points

  • Evaluation uses LoCoMo with the GPT-4o-mini and Qwen2.5-1.5B backbones. Results are reported in F1, BLEU-1 and ROUGE-2 across four categories: Multi-Hop, Temporal, Open-Domain and Single-Hop.
  • Input context is over 95% shorter than with long-context baselines and roughly 2–4× shorter than with A-Mem.
  • The graph layer is the most critical component. With GPT-4o-mini, removing it lowers Temporal F1 from 50.83 to 9.58 and Multi-Hop F1 from 26.00 to 13.97. Removing experience memory costs less, with Multi-Hop F1 dropping to 24.86.
  • Performance changes only mildly when the number of retrieved triples or text items is varied, so heavy hyperparameter tuning is not needed.
  • In a human evaluation of evidence support (25 questions per category, GPT-4o-mini), MemWeaver's retrieved knowledge was judged more supportive than A-Mem's in all four categories.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 2The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 works

🧭 Topics 1Research subjects and languages the work is about.Reasoning · 1,685 works

Cite

Cite as Juexiang Ye, Xue Li, Yang Xinyu, Chengkai Huang, Lanshun Nie, Lina Yao and Dechen Zhan. 2026. MemWeaver: Weaving Hybrid Memories for Traceable Long-Horizon Agentic Reasoning. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.630

@inproceedings{ye2026memweaver,
  title = {{MemWeaver: Weaving Hybrid Memories for Traceable Long-Horizon Agentic Reasoning}},
  author = {Juexiang Ye, Xue Li, Yang Xinyu, Chengkai Huang, Lanshun Nie, Lina Yao and Dechen Zhan},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.630},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.630}
}

References

35 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Sebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai, Eliza Rutherford, Katie Millican, George van den Driessche, Jean-Baptiste Lespiau, Bogdan Damoc, Aidan Clark, Diego de Las Casas, Aurelia Guy, Jacob Menick, Roman Ring, Tom Hennigan, Saffron Huang, Loren Maggiore, Chris Jones, Albin Cas… find
  2. Prateek Chhikara, Dev Khant, Saket Aryan, Taranjeet Singh, and Deshraj Yadav. 2025. Mem0: Building production-ready AI agents with scalable long-term memory. CoRR, abs/2504.19413.
  3. Aixin Liu DeepSeek-AI, Aoxue Mei, Bangcai Lin, Bing Xue, Bingxuan Wang, Bingzheng Xu, Bochao Wu, Bowei Zhang, Chaofan Lin, Chen Dong, and 1 others. 2025. Deepseek-v3. 2: Pushing the frontier of open large language models. arXiv preprint arXiv:2512.02556.
  4. Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, and Jonathan Larson. 2024. From local to global: A graph RAG approach to query-focused summarization. CoRR, abs/2404.16130. find
  5. Bernal Jimenez Gutierrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga, and Yu Su. 2024. Hipporag: Neurobiologically inspired long-term memory for large language models. In Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2024, NeurIPS 2024, V…
  6. Bernal Jiménez Gutiérrez, Yiheng Shu, Weijian Qi, Sizhe Zhou, and Yu Su. 2025. From RAG to memory: Non-parametric continual learning for large language models. In Forty-second International Conference on Machine Learning, ICML 2025, Vancouver, BC, Canada, July 13-19, 2025. OpenReview.net.
  7. Chengkai Huang, Junda Wu, Zhouhang Xie, Yu Xia, Rui Wang, Tong Yu, Subrata Mitra, Julian McAuley, and Lina Yao. 2025b. Pluralistic off-policy evaluation and alignment. arXiv preprint arXiv:2509.19333. — [arXiv:2509.19333](https://arxiv.org/abs/2509.19333) find
  8. Chengkai Huang, Yu Xia, Rui Wang, Kaige Xie, Tong Yu, Julian McAuley, and Lina Yao. 2025c. Embedding-informed adaptive retrieval-augmented generation of large language models. In Proceedings of the 31st International Conference on Computational Linguistics, pages 1403–1412.
  9. Aaron Hurst, Adam Lerer, Adam P. Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, Aleksander Madry, Alex Baker-Whitcomb, Alex Beutel, Alex Borzunov, Alex Carney, Alex Chow, Alex Kirillov, Alex Nichol, Alex Paino, and 79 others. 2024. Gpt-4o sy…
  10. Sagar Imambi, Kolla Bhanu Prakash, and GR Kanagachidambaresan. 2021. Pytorch. In Programming with TensorFlow: solution for edge computing applications, pages 87–104. Springer. find
  11. Shuguang Jiao, Chengkai Huang, Shuhan Qi, Xuan Wang, Yifan Li, and Lina Yao. 2026a. Doctor-rag: Failure-aware repair for agentic retrieval-augmented generation. arXiv preprint arXiv:2604.00865. — [arXiv:2604.00865](https://arxiv.org/abs/2604.00865) find
  12. Shuguang Jiao, Xinyu Xiao, Yunfan Wei, Shuhan Qi, Chengkai Huang, Quan Z. Sheng, and Lina Yao. 2026b. Prunerag: Confidence-guided query decomposition trees for efficient retrieval-augmented generation. In Proceedings of the ACM Web Conference 2026, WWW 2026, Dubai, United Arab Emirates, originally s… find
  13. Omkaresh Kulkarni and Adnan Burhanpurwala. 2024. A survey of advancements in dbscan clustering algorithms for big data. In 2024 3rd International conference on Power Electronics and IoT Applications in Renewable Energy and its Control (PARC), pages 106–111. IEEE. find
  14. Kuang-Huei Lee, Xinyun Chen, Hiroki Furuta, John F. Canny, and Ian Fischer. 2024. A human-inspired reading agent with gist memory of very long contexts. In Forty-first International Conference on Machine Learning, ICML 2024, Vienna, Austria, July 21-27, 2024. OpenReview.net.
  15. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-augmented generation for knowledge-intensive NLP tasks. In Advances in Neural Information Proc…
  16. Lei Liu, Xiaoyan Yang, Yue Shen, Binbin Hu, Zhiqiang Zhang, Jinjie Gu, and Guannan Zhang. 2023. Thinkin-memory: Recalling and post-thinking enable llms with long-term memory. CoRR, abs/2311.08719. — [arXiv:2311.08719](https://arxiv.org/abs/2311.08719) find
  17. Adyasha Maharana, Dong-Ho Lee, Sergey Tulyakov, Mohit Bansal, Francesco Barbieri, and Yuwei Fang. 2024. Evaluating very long-term conversational memory of LLM agents. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2024, Bangkok… find
  18. Siru Ouyang, Jun Yan, I-Hung Hsu, Yanfei Chen, Ke Jiang, Zifeng Wang, Rujun Han, Long T. Le, Samira Daruki, Xiangru Tang, Vishy Tirumalashetty, George Lee, Mahsan Rofouei, Hangfei Lin, Jiawei Han, Chen-Yu Lee, and Tomas Pfister. 2025. Reasoningbank: Scaling agent self-evolving with reasoning memory.… — [arXiv:2509.25140](https://arxiv.org/abs/2509.25140) find
  19. Charles Packer, Vivian Fang, Shishir G. Patil, Kevin Lin, Sarah Wooders, and Joseph E. Gonzalez. 2023. Memgpt: Towards llms as operating systems. CoRR, abs/2310.08560. find
  20. Joon Sung Park, Joseph C. O’Brien, Carrie Jun Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. 2023. Generative agents: Interactive simulacra of human behavior. In Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology, UIST 2023, San Francisco, CA,… find
  21. Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, EMNLP-IJCNLP 2019, Hong Kong, Ch…
  22. Erich Schubert, Jörg Sander, Martin Ester, Hans-Peter Kriegel, and Xiaowei Xu. 2017. DBSCAN revisited, revisited: Why and how you should (still) use DBSCAN. ACM Trans. Database Syst., 42(3):19:1– 19:21. find
  23. Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. 2023. Reflexion: language agents with verbal reinforcement learning. In Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems 2023, NeurIPS 2023, New Orleans,…
  24. Jiashuo Sun, Chengjin Xu, Lumingyuan Tang, Saizhuo Wang, Chen Lin, Yeyun Gong, Lionel M. Ni, Heung-Yeung Shum, and Jian Guo. 2024. Think-on-graph: Deep and responsible reasoning of large language model on knowledge graph. In The Twelfth International Conference on Learning Representations, ICLR 2024… find
  25. Llama Team. 2024. The llama 3 herd of models. CoRR, abs/2407.21783.
  26. Bing Wang, Xinnian Liang, Jian Yang, Hui Huang, Shuangzhi Wu, Peihao Wu, Lu Lu, Zejun Ma, and Zhoujun Li. 2023. Scm: Enhancing large language model with self-controlled memory framework. arXiv e-prints, pages arXiv–2304. find
  27. Shanglin Wu, Lihui Liu, Jinho D. Choi, and Kai Shu. 2025. Improving factuality in llms via inference-time knowledge graph construction. CoRR, abs/2509.03540. — [arXiv:2509.03540](https://arxiv.org/abs/2509.03540) find
  28. Wujiang Xu, Zujie Liang, Kai Mei, Hang Gao, Juntao Tan, and Yongfeng Zhang. 2025. A-MEM: agentic memory for LLM agents. CoRR, abs/2502.12110.
  29. An Yang, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoyan Huang, Jiandong Jiang, Jianhong Tu, Jianwei Zhang, Jingren Zhou, Junyang Lin, Kai Dang, Kexin Yang, Le Yu, Mei Li, Minmin Sun, Qin Zhu, Rui Men, Tao He, and 9 others. 2025. Qwen2.5-1m technical report. CoRR, abs/2501.15383.
  30. Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Bengio, William W. Cohen, Ruslan Salakhutdinov, and Christopher D. Manning. 2018. Hotpotqa: A dataset for diverse, explainable multi-hop question answering. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Brussel…
  31. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R. Narasimhan, and Yuan Cao. 2023. React: Synergizing reasoning and acting in language models. In The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023. OpenReview.net.
  32. Michihiro Yasunaga, Hongyu Ren, Antoine Bosselut, Percy Liang, and Jure Leskovec. 2021. QA-GNN: reasoning with language models and knowledge graphs for question answering. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Lang… find
  33. Guibin Zhang, Muxin Fu, Guancheng Wan, Miao Yu, Kun Wang, and Shuicheng Yan. 2025. G-memory: Tracing hierarchical memory for multi-agent systems. CoRR, abs/2506.07398. — [arXiv:2506.07398](https://arxiv.org/abs/2506.07398) find
  34. Lingfeng Zhong, Jia Wu, Qian Li, Hao Peng, and Xindong Wu. 2024a. A comprehensive survey on automatic knowledge graph construction. ACM Comput. Surv., 56(4):94:1–94:62. find
  35. Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. 2024b. Memorybank: Enhancing large language models with long-term memory. In Thirty-Eighth AAAI Conference on Artificial Intelligence, AAAI 2024, Thirty-Sixth Conference on Innovative Applications of Artificial Intelligence, IAAI 2024, F… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ye-2026-memweaver-weaving-hybrid-memories-traceable/