Elamur Research⌕
☰
home / papers / 2026 / Amory: Building Coherent Narrative-Driven Agent Memory through Agentic Reasoning

Amory: Building Coherent Narrative-Driven Agent Memory through Agentic Reasoning

Paper · 2026 · Yue Zhou, Xiaobo Guo, Belhassen Bayar et al.

Summary

Conversational agents that interact with users over weeks or months cannot afford to reread the whole dialogue history every turn. Most existing memory layers split dialogue into separate embeddings or graph nodes and fetch them in a RAG-like way, so memory construction stays shallow. The authors propose Amory, a working-memory framework grounded in cognitive science. Its heavy reasoning runs asynchronously, offline, where an LLM agent builds memory in three ways:

  • it binds new dialogue fragments into episodic narratives (characters, a headline plot, and a tree of subplots and timestamped fragments);
  • it consolidates those narratives guided by conversational momentum;
  • it moves peripheral facts into a triple-based semantic memory.

When a query arrives, an agentic retriever judges which narratives are coherent with the question, instead of ranking them by embedding similarity. On the LOCOMO long-term reasoning benchmark, Amory clearly beats earlier memory systems. Its answers are close to full-context reasoning in quality, at roughly half the response time. This suggests that richer offline memory construction can deliver both quality and speed for real-time agents.

Key points

  • Combining episodic and semantic memory gives up to +27.8% absolute over Mem0 on LOCOMO. Other baselines include RAG, HippoRAG, A-Mem, ReadAgent and Zep.
  • Inactive (momentum-aware) consolidation improves temporal reasoning compared with rapid consolidation or none.
  • The coherence-driven retriever covers more relevant memory than an embedding retriever, especially on multi-hop questions. At top-k = 2 it typically uses under 3.7% of the conversation (median compression above 96.3%).
  • For an agentic test, the authors converted 200 AgentIF instructions into one interleaved conversation of about 350,000 tokens. AgentIF itself contains 707 instructions from 50 applications.
  • On this task, Amory recovers 47.4% of the constraints. The closest baseline is ReadAgent at 36.8%, followed by A-Mem (27.8), Mem0 (25.1), HippoRAG (23.9) and Zep (21.3).

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

 Institutions

🏢AmazonSeattle · United States128 works · 41,697 citations🎓University of Illinois ChicagoChicago · United States70 works · 20,363 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 works

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

Cite

Cite as Yue Zhou, Xiaobo Guo, Belhassen Bayar and Srinivasan H. Sengamedu. 2026. Amory: Building Coherent Narrative-Driven Agent Memory through Agentic Reasoning. Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.eacl-long.183

@inproceedings{zhou2026amory,
  title = {{Amory: Building Coherent Narrative-Driven Agent Memory through Agentic Reasoning}},
  author = {Yue Zhou, Xiaobo Guo, Belhassen Bayar and Srinivasan H. Sengamedu},
  booktitle = {Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.eacl-long.183},
  url = {https://doi.org/10.18653/v1/2026.eacl-long.183}
}

References

21 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Tory S Anderson and Gatech Edu. 2015. Goal reasoning and narrative cognition. In Goal Reasoning: Papers from the ACS Workshop, page 1. 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. arXiv preprint arXiv:2504.19413.
  3. Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, Dasha Metropolitansky, Robert Osazuwa Ness, and Jonathan Larson. 2024. From local to global: A graph rag approach to query-focused summarization. arXiv preprint arXiv:2404.16130. find
  4. Daniel L Greenberg and Mieke Verfaellie. 2010. Interdependence of episodic and semantic memory: Evidence from neuropsychology. Journal of the International Neuropsychological society, 16(5):748–753. find
  5. Zirui Guo, Lianghao Xia, Yanhua Yu, Tu Ao, and Chao Huang. 2025. Lightrag: Simple and fast retrieval-augmented generation. Preprint, arXiv:2410.05779.
  6. Jie Huang and Kevin Chen-Chuan Chang. 2023. Towards reasoning in large language models: A survey. In Findings of the Association for Computational Linguistics: ACL 2023, pages 1049–1065, Toronto, Canada. Association for Computational Linguistics. \#\# Limitations Bernal Jimenez Gutierrez, Yiheng Shu,…
  7. While Amory demonstrates substantial improvements over existing memory frameworks, several Kuang-Huei Lee, Xinyun Chen, Hiroki Furuta, John Canny, and Ian Fischer. 2024. A human-inspired reading agent with gist memory of very long contexts. In Proceedings of the 41st International Conference on Mach…
  8. 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. 2021. Retrieval-augmented generation for knowledge-intensive nlp tasks. Preprint, arXiv:2005.11401.
  9. Zhiyu Li, Shichao Song, Hanyu Wang, Simin Niu, Ding Chen, Jiawei Yang, Chenyang Xi, Huayi Lai, Jihao Zhao, Yezhaohui Wang, Junpeng Ren, Zehao Lin, Jiahao Huo, Tianyi Chen, Kai Chen, Kehang Li, Zhiqiang Yin, Qingchen Yu, Bo Tang, and 3 others. 2025. Memos: An operating system for memory-augmented gen… — [arXiv:2505.22101](https://arxiv.org/abs/2505.22101) find
  10. 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), pages 13851– 1387… find
  11. Charles Packer, Sarah Wooders, Kevin Lin, Vivian Fang, Shishir G. Patil, Ion Stoica, and Joseph E. Gonzalez. 2024. Memgpt: Towards llms as operating systems. Preprint, arXiv:2310.08560. find
  12. Yunjia Qi, Hao Peng, Xiaozhi Wang, Amy Xin, Youfeng Liu, Bin Xu, Lei Hou, and Juanzi Li. 2025. Agentif: Benchmarking instruction following of large language models in agentic scenarios. Preprint, arXiv:2505.16944. — [arXiv:2505.16944](https://arxiv.org/abs/2505.16944) find
  13. Preston Rasmussen, Pavlo Paliychuk, Travis Beauvais, Jack Ryan, and Daniel Chalef. 2025. Zep: A temporal knowledge graph architecture for agent memory. Preprint, arXiv:2501.13956.
  14. Larry R Squire, Lisa Genzel, John T Wixted, and Richard G Morris. 2015. Memory consolidation. Cold Spring Harbor perspectives in biology, 7(8):a021766. find
  15. Endel Tulving and 1 others. 1972. Episodic and semantic memory. Organization of memory, 1(381-403):1. find
  16. Wujiang Xu, Kai Mei, Hang Gao, Juntao Tan, Zujie Liang, and Yongfeng Zhang. 2025. A-mem: Agentic memory for llm agents. arXiv preprint arXiv:2502.12110.
  17. Andrew P Yonelinas, Charan Ranganath, Arne D Ekstrom, and Brian J Wiltgen. 2019. A contextual binding theory of episodic memory: systems consolidation reconsidered. Nature Reviews Neuroscience, 20(6):364–375. find
  18. Guibin Zhang, Muxin Fu, Guancheng Wan, Miao Yu, Kun Wang, and Shuicheng Yan. 2025. G-memory: Tracing hierarchical memory for multi-agent systems. Preprint, arXiv:2506.07398. — [arXiv:2506.07398](https://arxiv.org/abs/2506.07398) find
  19. Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. 2023. Memorybank: Enhancing large language models with long-term memory. Preprint, arXiv:2305.10250. find
  20. Yue Zhou and Barbara Di Eugenio. 2025. Veracity bias and beyond: Uncovering LLMs’ hidden beliefs in problem-solving reasoning. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 21298–21310, Vienna, Austria. Association for Compu… find
  21. Yue Zhou, Yada Zhu, Diego Antognini, Yoon Kim, and Yang Zhang. 2024. Paraphrase and solve: Exploring and exploiting the impact of surface form on mathematical reasoning in large language models. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zhou-2026-amory-building-coherent-narrative-driven/