The DOI matches its Crossref record; 32 of the 32 references it cites were found in a registry or an index; 11 of the 17 sources its claims cite were checked for support.
A-Mem: Agentic Memory for LLM Agents
Open accessAbstract
This paper proposes a novel agentic memory system for LLM agents that can dynamically organize memories in an agentic way following Zettelkasten note-taking principles with dynamic indexing and linking.
Overview
CoChat Summary
TL;DR
A-MEM is an agentic memory system for LLM agents that, inspired by the Zettelkasten method, builds structured notes with LLM-generated keywords, tags, and contextual descriptions, then autonomously links and evolves them as new memories arrive. Across six foundation models on LoCoMo and DialSim, it outperforms LoCoMo, ReadAgent, MemoryBank, and MemGPT while using far fewer tokens.
Key findings
- On LoCoMo with GPT-4o-mini, A-MEM reaches 27.02 F1 / 20.09 BLEU-1 on Multi Hop and 45.85 / 36.67 on Temporal, versus MemGPT's 26.65 / 17.72 and 25.52 / 19.44; A-MEM's average ranking is 1.2 versus 2.4 for both LoCoMo and MemGPT.
- On DialSim, A-MEM scores F1 3.45, BLEU-1 3.37, ROUGE-L 3.54, ROUGE-2 3.60, METEOR 2.05, SBERT 19.51, versus LoCoMo (2.55, 3.13, 2.75, 0.90, 1.64, 15.76) and MemGPT (1.18, 1.07, 0.96, 0.42, 0.95, 8.54) — a 35% F1 improvement over LoCoMo and 192% over MemGPT.
- Ablation with GPT-4o-mini: removing both link generation and memory evolution drops Multi Hop F1 to 9.65; removing only memory evolution gives 21.35; full A-MEM gives 27.02.
- Cost/scaling: ~1,200 tokens per memory operation (85–93% reduction vs. ~16,900 for LoCoMo/MemGPT), under $0.0003 per operation, 5.4 s average with GPT-4o-mini and 1.1 s with local Llama 3.2 1B; retrieval time grows only from 0.31 µs to 3.70 µs from 1,000 to 1,000,000 memories.
Methods
- Design: each memory note is `{content, timestamp, keywords, tags, context, embedding, links}`; an LLM generates keywords/tags/context, a text encoder (all-minilm-l6-v2) produces embeddings, top-k nearest neighbors are retrieved by cosine similarity, and an LLM decides link generation and memory evolution (updating neighbors' context, keywords, tags).
- Data: LoCoMo (7,512 QA pairs, dialogues averaging 9K tokens over up to 35 sessions; single-hop, multi-hop, temporal, open-domain, adversarial) and DialSim (multi-party dialogues from Friends, The Big Bang Theory, The Office; 1,300 sessions, ~350,000 tokens, >1,000 questions per session).
- Evaluation: F1 and BLEU-1 as primary metrics, plus ROUGE-L, ROUGE-2, METEOR, and SBERT similarity; baselines LoCoMo, ReadAgent, MemoryBank, MemGPT; six foundation models (GPT-4o-mini, GPT-4o, Qwen2.5-1.5B/3B, Llama 3.2 1B/3B) plus DeepSeek-R1-32B, Claude 3.0 Haiku, and Claude 3.5 Haiku in the appendix; top-k retrieval with k=10 default and category-specific tuning.
Limitations
- Stated by the paper: memory organization quality may still depend on the underlying LLM's capabilities, and different LLMs may generate different contextual descriptions or connections; the implementation is text-only, with multimodal (image/audio) extension left to future work.
- Evident from the text: no error bars or significance tests are reported (the checklist answers "No" to statistical significance, citing API cost); the main-text results are reported for a single dataset family of long-term conversation QA, and the paper's own checklist notes broader societal impacts are not discussed.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
5 Valid · 0 Not valid · 3 Not checked · 55 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“These systems require agent developers to predefine memory storage structures, specify storage points within the workflow, and establish retrieval timing.”
What the sources say
- Partly supportedcites [25] [25] Charles Packer, Sarah Wooders, Kevin Lin, Vivian Fang, Shishir G Patil, Ion Stoica, and Joseph E Gonzalez. Memgpt: Towards llms as operating systems. arXiv preprint arXiv:2310.08560, 2023. sourceExperimental — accuracy not yet measured
Section 3.2.1 (p. 6) says the fixed-context baselines fetch documents independently of LLM inference, supporting the idea that retrieval is set by the workflow rather than chosen by the agent. Section 2.3 (p. 3) contrasts this with MemGPT’s self-directed memory edits and retrieval. The paper does not establish that existing memory systems generally require developers to predefine storage structures or specify storage points. (LLM, conf=0.78)
- Could not verifycites [39] [39] Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. Memorybank: Enhancing large language models with long-term memory. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 19724–19731, 2024. 12 Contents sourceExperimental — accuracy not yet measured
Sections 2.1 and 2.2 describe MemoryBank’s storage structure and retrieval mechanism, but do not say that existing systems require agent developers to predefine structures, choose workflow storage points, or set retrieval timing. Searches for those developer requirements found no passage addressing them. (LLM, conf=0.73)
- Not matchedcites [28] [28] Aymeric Roucher, Albert Villanova del Moral, Thomas Wolf, Leandro von Werra, and Erik Kaunismäki. 'smolagents': a smol library to build great agentic systems. https://github. com/huggingface/smolagents, 2025.
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Could not verifycites [21] [21] Zhiwei Liu, Weiran Yao, Jianguo Zhang, Liangwei Yang, Zuxin Liu, Juntao Tan, Prafulla K Choubey, Tian Lan, Jason Wu, Huan Wang, et al. Agentlite: A lightweight library for building and advancing task-oriented llm agent system. arXiv preprint arXiv:2402.15538, 2024. sourceExperimental — accuracy not yet measured
Section 2.1 describes when AgentLite retrieves an action-observation history and stores an observation; the Figure 1 description in Section 6 also describes its memory workflow. Neither passage says that existing memory systems require developers to predefine storage structures, storage points, or retrieval timing. Searches for those requirements found no passage addressing them. (LLM, conf=0.72)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Existing memory systems [25, 39, 28, 21] for LLM agents provide basic memory storage functionality.”
What the sources say
- Supportedcites [25] [25] Charles Packer, Sarah Wooders, Kevin Lin, Vivian Fang, Shishir G Patil, Ion Stoica, and Joseph E Gonzalez. Memgpt: Towards llms as operating systems. arXiv preprint arXiv:2310.08560, 2023. source
MemGPT is an LLM-agent memory system that stores information outside the context window. Section 2 (p. 2) describes its external context and memory-management functions, and the Figure 3 description (p. 3) identifies archival and recall storage databases. The claim does not say these systems provide only basic storage. (LLM, conf=0.88)
- Supportedcites [39] [39] Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. Memorybank: Enhancing large language models with long-term memory. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 19724–19731, 2024. 12 Contents source
MemoryBank is an LLM memory system with a memory-storage component. Section 2 describes that component as its primary data repository, and §2.1 says it stores conversation records, event summaries, and user information. This supports citing MemoryBank as an example of a system that provides memory storage functionality. (LLM, conf=0.91)
- Not matchedcites [28] [28] Aymeric Roucher, Albert Villanova del Moral, Thomas Wolf, Leandro von Werra, and Erik Kaunismäki. 'smolagents': a smol library to build great agentic systems. https://github. com/huggingface/smolagents, 2025.
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Supportedcites [21] [21] Zhiwei Liu, Weiran Yao, Jianguo Zhang, Liangwei Yang, Zuxin Liu, Juntao Tan, Prafulla K Choubey, Tian Lan, Jason Wu, Huan Wang, et al. Agentlite: A lightweight library for building and advancing task-oriented llm agent system. arXiv preprint arXiv:2402.15538, 2024. source
Section 2.1, “Individual Agent” (page not available), describes AgentLite’s Memory component as storing an agent’s action-observation chains and saving new actions and observations during execution. This supports citing AgentLite as an example of an LLM-agent system with basic memory storage functionality. (LLM, conf=0.90)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“While previous datasets contain dialogues with around 1K tokens over 4-5 sessions, LoCoMo features much longer conversations averaging 9K tokens spanning up to 35 sessions”
§ 4 Experiment > 4.1 Dataset and Evaluation
What the sources say
- Supportedcites [22] [22] Adyasha Maharana, Dong-Ho Lee, Sergey Tulyakov, Mohit Bansal, Francesco Barbieri, and Yuwei Fang. Evaluating very long-term conversational memory of llm agents. arXiv preprint arXiv:2402.17753, 2024. source
The Introduction (p. 2) reports that LoCoMo conversations average 9K tokens and span up to 35 sessions. Its comparison table (p. 2) lists earlier multi-session datasets at about 1,055–1,226 tokens across four or five sessions, supporting the stated contrast. (LLM, conf=0.94)
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“T-SNE Visualization of Memory Embeddings Showing More Organized Distribution with A-MEM (blue) Compared to Base Memory (red) Across Different Dialogues.”
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“with each memory operation costing less than $0.0003 when using commercial API services”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
25 more claims
Create a free account to see all 63 claims and every source verdict.
- Probably not a substantial claim. 16
- It may state two separate claims as one. 20
- The quote only partly states this claim. 2
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source doesn’t support the claim as written.
In the citing paper
“Relative to the O (1) dictionary lookup used by all prior systems [4], [5], [6], [7], [8], [9], lineage-gated retrieval trades at most 13 . 82 µ s in absolute terms for the safety guarantee of Theorem 1-a tradeoff Experiments 1-2 show buys elimination of an 18 . 8 -25 . 5% structural leak rate at a 13 -14 percentage-point reuse cost.”
From Lineage-Aware Memory Governance: A Derivation-Gated Framework for Privacy-Preserving Column-Level Access Control in Enterprise AI Agents (2026) · cites this paper as [6]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“This emerging class of systems extends the capabilities of traditional AI Agents by enabling multiple intelligent entities to collaboratively pursue goals through structured communication [144-146], shared memory [147,148], and dynamic role assignment [22].”
From AI Agents vs. Agentic AI: A Conceptual taxonomy, applications and challenges (2026) · cites this paper as [147]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“from Wang et al. [55], Xu et al. [45] build memory to improve long context performance for specific workflows which suggest that diverse evaluation workflows are needed for better testing of generalizable memory components.”
From MEMTRACK: Evaluating Long-Term Memory and State Tracking in Multi-Platform Dynamic Agent Environments (2025) · cites this paper as [45]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Some other approaches introduce external memory modules ( e.g., summarizers or retrievers) [63, 29, 13, 57], but these are typically trained separately and cannot be optimized end-to-end with the agent's policy. This also introduces additional engineering overhead, as engineers must manage and integrate two separate models.”
From MEM1: Learning to Synergize Memory and Reasoning for Efficient Long-Horizon Agents (2025) · cites this paper as [57]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Some others, such as HiAgent[5], A-Mem[23], MemAgent [28] use hierarchical designs to separate working buffers from episodic and long-term stores to relieve context pressure and improve persistence.”
From Learning Hierarchical Procedural Memory for LLM Agents through Bayesian Selection and Contrastive Refinement (2025) · cites this paper as [23]
Experimental — accuracy not yet measured
9 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ APPENDIX
“Using GPT-4o-mini, A-MEM achieves a ROUGE-L score of 44.27 in Multi-Hop tasks, more than doubling LoComo's 18.09.”
§ APPENDIX
“LOCOMO 9.64 23.92 2.01 18.09 3.40 11.58 26.48 40.20 60.46 69.59 READAGENT 2.47 9.45 0.95 13.12 0.55 5.76 2.99 9.92 6.66 9.79 MEMORYBANK 1.18 5.43 0.52 9.64 0.97 5.77 1.64 6.63 4.55 7.35 MEMGPT 10.58 25.60 4.76 25.22 0.76 9.14 28.44 42.24 36.62 43.75 A-MEM 10.61 25.86 21.39 44.27 3.42 12.09 29.50 45.18 42.62 50.04”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ APPENDIX
“The Multi-Hop category showcases particularly striking results, where Qwen2.5-15b with A-MEM achieves a ROUGE-L score of 27.23, dramatically surpassing LoComo's 4.68 and ReadAgent's 2.81 - representing a nearly six-fold improvement.”
§ APPENDIX
“Multi Hop Temporal Open Domain Single Hop Adversial RGE-2 RGE-L RGE-2 RGE-L RGE-2 RGE-L RGE-2 RGE-L RGE-2 RGE-L”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ APPENDIX
“This significant advantage maintains consistency across other metrics, with METEOR scores of 23.43 versus 7.61 and SBERT scores of 70.49 versus 52.30.”
§ APPENDIX
“Multi Hop Temporal Open Domain Single Hop Adversial ME SBERT ME SBERT ME SBERT ME SBERT ME SBERT”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ 4 Experiment > 4.3 Empricial Results
“our A-MEM demonstrates superior performance in Multi-Hop tasks achieves at least two times better performance”
§ 4 Experiment > 4.1 Dataset and Evaluation
“MEMGPT 26.65 17.72 25.52 19.44 9.15 7.44 41.04 34.34 43.29 42.73 2.4 2.4 16,977 A-MEM 27.02 20.09 45.85 36.67 12.14 12.00 44.65 37.06 50.03 49.47 1.2 1.2 2,520”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 10 20 30 40 50
“44.55 25.43 28.31 32.12 34.32 37.02”
§ 4 Experiment > 4.1 Dataset and Evaluation
“A-MEM 27.02 20.09 45.85 36.67 12.14 12.00 44.65 37.06 50.03 49.47 1.2 1.2 2,520”
Citations
Cites
The publisher hasn’t deposited this paper’s reference list.
Research works printed but not linked: 4
- [1][1] Sönke Ahrens. How to Take Smart Notes: One Simple Technique to Boost Writing, Learning and Thinking. Amazon, 2017. Second Edition.
- [5][5] Satanjeev Banerjee and Alon Lavie. Meteor: An automatic metric for mt evaluation with improved correlation with human judgments. In Proceedings of the acl workshop on intrinsic and extrinsic evaluation measures for machine translation and/or summarization, pages 65–72, 2005.
- [18][18] Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, et al. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems, 33:9459–9474, 2020.
- [19][19] Chin-Yew Lin. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pages 74–81, 2004.
Cited by
A Survey of Large Language Models
Zhao, W. X., Zhou, K., Li, J., Tang, T., Dong, Z., Hou, Y., Zhang, B., Min, Y., Zhang, J., Liu, P., et al. (2026) · Frontiers of Computer Science · doi:10.1007/s11704-026-60308-3 · Cited by 2,093
AI Agents vs. Agentic AI: A Conceptual taxonomy, applications and challenges
Sapkota, R., Roumeliotis, K. I., & Karkee, M. (2026) · Information Fusion · doi:10.1016/j.inffus.2025.103599 · Cited by 671
“This emerging class of systems extends the capabilities of traditional AI Agents by enabling multiple intelligent entities to collaboratively pursue goals through structured communication [146]–[148], shared memory [149], [150], and dynamic role assignment [21].”
Memory in the Age of AI Agents
Hu, Y., Liu, S., Yue, Y., Zhang, G., Liu, B., Zhu, F., Lin, J., Guo, H., Dou, S., Xi, Z., et al. (2025) · arXiv · doi:10.48550/arxiv.2512.13564 · Cited by 389
Self-citationAgentic Context Engineering: Evolving Contexts for Self-Improving Language Models
Zhang, Q., Hu, C., Upasani, S., Ma, B., Hong, F., Kamanuru, V., Rainton, J., Wu, C., Ji, M., Li, H., et al. (2025) · arXiv · doi:10.48550/arxiv.2510.04618 · Cited by 370
“A-MEM (Xu et al., 2025) introduces a dynamically organized memory system inspired by the Zettelkasten method: each stored memory is annotated with structured attributes ( e.g., tags, keywords, contextual descriptions) and automatically linked to relevant past entries, while existing entries are…”
InfluentialYan, S., Yang, X., Huang, Z., Nie, E., Ding, Z., Li, Z., Ma, X., Bi, J., Kersting, K., Pan, J. Z., et al. (2025) · arXiv · doi:10.48550/arxiv.2508.19828 · Cited by 297
“…relevant segments for answering questions, serving as the benchmark baseline for long-range, multi-session conversation reasoning; (2) A-Mem (Xu et al., 2025), a dynamic agen-tic memory system that creates, links, and updates structured memories to enhance reasoning across sessions; (3)…”
ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory
Ouyang, S., Yan, J., Hsu, I.-H., Chen, Y., Jiang, K., Wang, Z., Han, R., Le, L. T., Daruki, S., Tang, X., et al. (2025) · arXiv · doi:10.48550/arxiv.2509.25140 · Cited by 295
AI Agents vs. Agentic AI: A Conceptual Taxonomy, Applications and Challenges
Sapkota, R., Roumeliotis, K. I., & Karkee, M. (2025) · arXiv · doi:10.48550/arxiv.2505.10468 · Cited by 295
“This emerging class of systems extends the capabilities of traditional AI Agents by enabling multiple intelligent entities to collaboratively pursue goals through structured communication [146]–[148], shared memory [149], [150], and dynamic role assignment [21].”
A Survey of Context Engineering for Large Language Models
Mei, L., Yao, J., Ge, Y., Wang, Y., Bi, B., Cai, Y., Liu, J., Li, M., Li, Z.-Z., Zhang, D., et al. (2025) · arXiv · doi:10.48550/arxiv.2507.13334 · Cited by 281
A Survey of Reinforcement Learning for Large Reasoning ModelsPreprint
Zhang, K., Zuo, Y., He, B., Sun, Y., Liu, R., Jiang, C., Fan, Y., Tian, K., Jia, G., Li, P., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2509.08827 · Cited by 270
Self-citationThe landscape of agentic reinforcement learning for LLMs: a survey
Zhang, G., Geng, H., Yu, X., Yin, Z., Zhang, Z., Tan, Z., Zhou, H., Li, Z., Xue, X., Li, Y., et al. (2025) · arXiv · doi:10.48550/arxiv.2509.02547 · Cited by 229
MEM1: Learning to Synergize Memory and Reasoning for Efficient Long-Horizon Agents
Zhou, Z., Qu, A., Wu, Z., Kim, S., Prakash, A., Rus, D., Zhao, J., Low, B. K. H., & Liang, P. P. (2025) · arXiv · doi:10.48550/arxiv.2506.15841 · Cited by 218
“To mitigate these issues, recent studies have proposed external memory frameworks, including retrieval-augmented generation and summarization modules [63, 29, 13, 57].”
MIRIX: Multi-Agent Memory System for LLM-Based Agents
Wang, Y., & Chen, X. (2025) · arXiv · doi:10.48550/arxiv.2507.07957 · Cited by 201
“We compare MIRIX against the following baselines: A-Mem [35]: A memory system that builds Zettelkasten-style knowledge graphs from user-agent interactions, dynamically linking notes using embedding similarity and LLM reasoning.”
AI Agents vs. Agentic AI: A Conceptual Taxonomy, Applications and Challenges
Sapkota, R., Roumeliotis, K. I., & Karkee, M. (2025) · SuperIntelligence - Robotics - Safety & Alignment · doi:10.70777/si.v2i3.15161 · Cited by 186
A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and DeploymentPreprint
Wang, K., Zhang, G., Zhou, Z., Wu, J., Yu, M., Zhao, S., Yin, C., Fu, J., Yan, Y., Luo, H., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.15585 · Cited by 196
Self-citationReSum: Unlocking Long-Horizon Search Intelligence via Context SummarizationPreprint
Wu, X., Li, K., Zhao, Y., Zhang, L., Ou, L., Yin, H., Zhang, Z., Yu, X., Zhang, D., Jiang, Y., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2509.13313 · Cited by 169
A Survey on Large Language Model-Based Game Agents
Hu, S., Huang, T., Liu, G., Kompella, R. R., Ilhan, F., Tekin, S. F., Xu, Y., Yahn, Z., & Liu, L. (2024) · arXiv · doi:10.48550/arxiv.2404.02039 · Cited by 165
“Episodic memory GenerativeAgents [59], Xu et al. [28], ProAgent [61], etc.”
MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic MemoryPreprint
Zhang, S., Wang, J., Zhou, R., Liao, J., Feng, Y., Li, Z., Zheng, Y., Zhang, W., Wen, Y., Li, Z., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2601.03192 · Cited by 171
“More recent frameworks have system-atized the memory lifecycle, focusing on unified storage and structured indexing for complex tasks (Li et al., 2025b; Xu et al., 2025; Huang et al., 2025; Ye, 2025).”
AutoSkill: Experience-Driven Lifelong Learning via Skill Self-EvolutionPreprint
Yang, Y., Li, J., Pan, Q., Zhan, B., Cai, Y., Du, L., Zhou, J., Chen, K., Chen, Q., Li, X., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.01145 · Cited by 146
“More recent frameworks, including Mem0 [12], A MEM [18], and MemInsight [19], together with corresponding surveys [35], further systematize memory mechanisms for LLM based agents.”
AgentFold: Long-Horizon Web Agents with Proactive Context ManagementPreprint
Ye, R., Zhang, Z., Li, K., Yin, H., Tao, Z., Zhao, Y., Su, L., Zhang, L., Qiao, Z., Wang, X., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.24699 · Cited by 150
SkillX: Automatically Constructing Skill Knowledge Bases for AgentsPreprint
Wang, C., Yu, Z., Xie, X., Yao, W., Fang, R., Qiao, S., Cao, K., Zheng, G., Qi, X., Zhang, P., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2604.04804 · Cited by 119
Seeing, Listening, Remembering, and Reasoning: A Multimodal Agent with Long-Term Memory
Long, L., He, Y., Ye, W., Pan, Y., Lin, Y., Li, H., Zhao, J., & Li, W. (2025) · arXiv · doi:10.48550/arxiv.2508.09736 · Cited by 102
“Beyond raw data, some methods incorporate summaries [14, 21, 44, 59], latent embeddings [6, 28, 40, 56], or structured knowledge representations [33, 50].”
MoralBench: Moral Evaluation of LLMs
Ji, J., Chen, Y., Jin, M., Xu, W., Hua, W., & Zhang, Y. (2025) · ACM SIGKDD Explorations Newsletter · doi:10.1145/3748239.3748246 · Cited by 148
“…vast amounts of data and sophisticated algorithms, have demonstrated capabilities in understanding and generating human-like text [1; 2; 3; 4; 5; 6; 7], opening up new possibilities for applications ranging from automated programming [8; 9; 10] to complex decision-making systems [11; 12; 13; 14;…”
Self-citationMemGen: Weaving Generative Latent Memory for Self-Evolving AgentsPreprint
Zhang, G., Fu, M., & Yan, S. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2509.24704 · Cited by 113
Wei, T., Li, T.-W., Liu, Z., Ning, X., Yang, Z., Zou, J., Zeng, Z., Qiu, R., Lin, X., Fu, D., et al. (2026) · Transactions on machine learning research · doi:10.48550/arxiv.2601.12538 · Cited by 101
Self-citationA Survey of LLM-based Deep Search Agents: Paradigm, Optimization, Evaluation, and ChallengesPreprint
Xi, Y., Lin, J., Xiao, Y., Zhou, Z., Shan, R., Gao, T., Zhu, J., Liu, W., Yu, Y. H., & Zhang, W. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2508.05668 · Cited by 89
“Agentic search can extract queries from complex and ambiguous user intentions and then conduct deep searches within the agent’s memory to retrieve highly pertinent information (Xu et al., 2025b; Ocker et al., 2025; Tan et al., 2024b).”
A Survey of Vibe Coding with Large Language ModelsPreprint
Ge, Y., Mei, L., Duan, Z., Li, T., Zheng, Y., Wang, Y., Wang, L., Yao, J., Liu, T., Cai, Y., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.12399 · Cited by 96
A Survey of Agent Memory in the Second Half: Towards Self-Evolving and Long-Horizon Agents
Huang, W.-C., Zhang, W., Liang, Y., Bei, Y., Chen, Y., Feng, T., Pan, X., Tan, Z., Wang, Y., Wei, T., et al. (2026) · arXiv · doi:10.48550/arxiv.2602.06052 · Cited by 70
Self-citationSkill-Pro: Learning Reusable Skills from Experience via Non-Parametric PPO for LLM AgentsPreprint
Mi, Q., Ma, Z., Yang, M., Li, H., Wang, Y., Zhang, H., & Wang, J. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.01869 · Cited by 82
LLM-empowered knowledge graph construction: A surveyPreprint
Bian, H. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.20345 · Cited by 68
“Frameworks such as A-MEM (Xu et al., 2025) model memory as interconnected “notes” enriched with contextual metadata, enabling continual reorganization and growth.”
Hindsight is 20/20: Building Agent Memory that Retains, Recalls, and ReflectsPreprint
Latimer, C., Boschi, N., Neeser, A., Bartholomew, C., Srivastava, G., Wang, X., & Ramakrishnan, N. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.12818 · Cited by 73
“A-Mem (Xu et al., 2025) uses the Zettelkasten method to create atomic notes with LLM-generated links that evolve over time, but treats all memory uniformly without separating facts from opinions.”
MemInsight: Autonomous Memory Augmentation for LLM Agents
Salama, R., Cai, J. C., Yuan, M., Currey, A., Sunkara, M., Zhang, Y., & Benajiba, Y. (2025) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2025.emnlp-main.1683 · Cited by 116
Self-citationA Survey of LLM-based Software Repair: Taxonomies, Design Paradigms, and Applications
Yang, B., Cai, Z., Liu, F., Le, B., Zhang, L., Bissyandé, T. F., Liu, Y., & Tian, H. (2026) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3845995 · Cited by 65
Scaling LLM Multi-turn RL with End-to-end Summarization-based Context ManagementPreprint
Miao, L., Sun, W., Du, W., Ling, Z., Yao, X., Liu, K., & Chen, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.06727 · Cited by 60
“Now the list of the number of daily visitors to the firefly habitat is [18 , 29 , 46 , 14 , 13 , 17 , 31 , 4 , 8 , 15 , 34 , 17 , 25 , 17 , 24 , 48 , 43 , 33 , 36 , 36 , 7 , 38 , 26 , 6 , 49 , 48 , 22 , 9 , 33 , 30] .”
Deep Research: A Systematic SurveyPreprint
Shi, Z., Chen, Y., Li, H., Sun, W., Ni, S., Lyu, Y., Fan, R.-Z., Jin, B., Weng, Y., Zhu, M., et al. (2025) · Preprints.org · doi:10.20944/preprints202511.2077.v1 · Cited by 62
Self-citationIn Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents
Tan, Z., Yan, J., Hsu, I.-H., Han, R., Wang, Z., Le, L. T., Song, Y., Chen, Y., Palangi, H., Lee, G., et al. (2025) · arXiv · Cited by 54
A Survey of Personalization: From RAG to Agent
Li, X., Jia, P., Xu, D., Wen, Y., Zhang, Y., Zhang, W., Wang, W., Wang, Y., Du, Z., Li, X., et al. (2026) · ACM Transactions on Information Systems · doi:10.1145/3802586 · Cited by 63
“Furthermore, Matryoshka [58] generates user preference summaries by dynamically retrieving and synthesizing historical data.”
Self-citationVision-as-Inverse-Graphics Agent via Interleaved Multimodal Reasoning
Yin, S., Ge, J., Wang, Z. Z., Wang, C., Li, X., Black, M. J., Darrell, T., Kanazawa, A., & Feng, H. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37464-6_30 · Cited by 69
A Survey of Personalization: From RAG to Agent
Li, X., Jia, P., Xu, D., Wen, Y., Zhang, Y., Zhang, W., Wang, W., Wang, Y., Du, Z., Li, X., et al. (2025) · arXiv · doi:10.48550/arxiv.2504.10147 · Cited by 52
“Furthermore, Matryoshka [58] generates user preference summaries by dynamically retrieving and synthesizing historical data.”
Self-citationFrom language to action: a review of large language models as autonomous agents and tool users
Chowa, S. S., Alvi, R., Rahman, S. S., Rahman, M. A., Raiaan, M. A. K., Islam, M. R., Hussain, M., & Azam, S. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-025-11471-9 · Cited by 50
“The profiling module defines an agent’s operational persona (e.g., developer, advisor, or task-specific role), conditioning its behavior and role policy through static profiles defined by experts or dynamic generative mechanisms [21, 138].”
ExpeRepair: Dual-Memory Enhanced LLM-Based Repository-Level Program Repair
Mu, F., Wang, J., Shi, L., Wang, S., LI, S.-L., & Wang, Q. (2026) · Proceedings of the ACM on software engineering. · doi:10.1145/3808181 · Cited by 49
Adaptation of Agentic AI: A Survey of Post-Training, Memory, and SkillsPreprint
Jiang, P., Lin, J., Shi, Z., Wang, Z., He, L., Wu, Y., Zhong, M., Song, P., Zhang, Q., Wang, H., et al. (2025) · arXiv (Cornell University) · doi:10.4230/lipics.esem.2026.62 · Cited by 52
Self-citationSGMem: Sentence Graph Memory for Long-Term Conversational AgentsPreprint
Wu, Y., Zhang, Y., Liang, S., & Liu, Y. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2509.21212 · Cited by 49
“Existing methods include MemoryBank (Zhong et al., 2024), which hierarchically summarizes events and aggregates personality insights; LD-Agent (Li et al., 2024), which separates long-and short-term memory banks for event summaries and contextual dialogue while updating user personas; RMM (Tan et al., 2025), which refl…”
Self-citationOAgents: An Empirical Study of Building Effective Agents
Zhu, H., Qin, T., Zhu, K., Huang, H., Guan, Y., Xia, J., Li, H., Yao, Y., Wang, N., Liu, P., et al. (2025) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2025.findings-emnlp.720 · Cited by 53
Dynamic Dual-Granularity Skill Bank for Agentic RL
Tu, S., Xu, C., Zhang, Q., Zhang, Y., Lan, X., Li, L., Li, D., & Zhao, D. (2026) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/2026.emnlp-main.491 · Cited by 44
Self-citationYang, J., Liu, X., Lv, W., Deng, K., Guo, S., Jing, L., Li, Y., Liu, S., Luo, X., Luo, Y., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2511.18538 · Cited by 47
Massive Values in Self-Attention Modules are the Key to Contextual Knowledge UnderstandingPreprint
Jin, M., Mei, K., Xu, W., Sun, M., Tang, R., Du, M., Liu, Z., & Zhang, Y. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2502.01563 · Cited by 69
“Following existing benchmarks, we choose the commonly-used topics covering Sports, Arts, Technology and Celebrity to feed into the LLM to generate factual knowledge QA related to these topics (Hu et al., 2023; Xu et al., 2025a).”
Self-citationA Survey on Agentic Multimodal Large Language ModelsPreprint
Yao, H., Zhang, R., Huang, J., Zhang, J., Wang, Y., Fang, B., Zhu, R., Jing, Y., Liu, S., Li, G., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.10991 · Cited by 48
“Although recent studies have explored agentic memory [21, 22, 88, 93], most of these works have focused primarily on the language modality, with limited exploration of multimodal settings.”
InfluentialLiCoMemory: Lightweight and Cognitive Agentic Memory for Efficient Long-Term Reasoning
Huang, Z., Tian, Z., Guo, Q., Zhang, F., Zhou, Y., Jiang, D., Xie, Z., & Zhou, X. (2026) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2026.findings-acl.1835 · Cited by 46
Evaluating Long-Horizon Memory for Multi-Party Collaborative Dialogues
Hu, C., Li, T., Gao, X., Chen, H., Bai, Y., Xu, D., Lin, T., Li, X., Han, Y., Pei, J., et al. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770855.3817589 · Cited by 45
“Together, these contributions position EverMemBench as a concrete step toward realistic evaluation of LLM memory and a cornerstone benchmark for developing next-generation LLMs that reason over time, roles, and collaborative interaction structure.”
Lightweight LLM Agent Memory with Small Language Models
Zhang, J., Zhang, C., Chen, S., Huang, Z., Zheng, P., Wang, Z., Guo, P., Mo, F., Bae, S.-H., Zou, J., et al. (2026) · Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2026.acl-long.588 · Cited by 35
Useful Memories Become Faulty When Continuously Updated by LLMs
Zhang, D., Lin, Y., Wu, Z., Sun, Y., Li, B., Li, D., & Peng, H. (2026) · arXiv · doi:10.48550/arxiv.2605.12978 · Cited by 35
Maragheh, R. Y., & Deldjoo, Y. (2026) · ACM Transactions on Recommender Systems · doi:10.1145/3845988 · Cited by 47
From Recall to Forgetting: Benchmarking Long-Term Memory for Personalized Agents
Uddin, M. N., Shubham, K., Blanco, E., Baral, C., & Wang, G. (2026) · arXiv · doi:10.48550/arxiv.2604.20006 · Cited by 32
“For example, in the monthly setting, Memo-ryOS (112.8) and A-Mem (112.0) both outperform Nemori (105.4) under memory presence accuracy, but Nemori receives a much smaller forgetting-based reduction (15.4 vs. 28.4 and 29.5), yielding a higher final aggregated FAMA score and moving it ahead of both systems in the final…”
InfluentialVibe Coding vs. Agentic Coding: Fundamentals and Practical Implications of Agentic AI
Sapkota, R., Roumeliotis, K. I., & Karkee, M. (2025) · arXiv · doi:10.48550/arxiv.2505.19443 · Cited by 32
“A critical feature is the presence of memory and context management [113], [114], facilitating persistent state tracking across multi-step workflows.”
O-Mem: Omni Memory System for Personalized, Long Horizon, Self-Evolving AgentsPreprint
Wang, P., Tian, M., Li, J. X., Yuan, L., Wang, Y., Chen, Q., Wang, T., Lu, Z., Ma, J., Jiang, Y. E., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2511.13593 · Cited by 34
“Agentic Memory [39] organizes interactions into distinct groups according to their semantic similarity, while Mem0 [5] extracts meaningful content from messages and stores the extracted information independently to support future retrieval.”
InfluentialAutonomous Agents for Scientific Discovery: Orchestrating Scientists, Language, Code, and PhysicsPreprint
Zhou, L., Ling, H., Fu, C., Huang, Y., Sun, M., Yu, W., Wang, X., Li, X., Su, X., Zhang, J., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.09901 · Cited by 37
SkillOps: Managing LLM Agent Skill Libraries as Self-Maintaining Software Ecosystems
Song, X., Pu, H., & Zhao, L. (2026) · arXiv · doi:10.48550/arxiv.2605.13716
Self-Improvements in Modern Agentic Systems: A SurveyPreprint
Ren, Z., Chen, Y., Guo, D., Rong, G., Li, T., Xiong, R. B., Lan, Q., Wang, W., Nanbo, L., Yang, Y., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2607.13104 · Cited by 36
Graph-Augmented Large Language Model Agents: Current Progress and Future Prospects
Liu, Y., Zhang, G., Wang, K., Li, S., Pan, S., & An, B. (2026) · IEEE Intelligent Systems · doi:10.1109/mis.2025.3642667 · Cited by 49
“As shown in Figure 5, in a tool graph, each node denotes an available tool for agents, and each edge models the functional dependencies and/or compatibility between tools.”
Look Back to Reason Forward: Revisitable Memory for Long-Context LLM AgentsPreprint
Shi, Y., Chen, Y., Wang, S., Li, S., Cai, H., Gu, Q., Wang, X., & Zhang, A. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2509.23040 · Cited by 33
“More recent approaches explore a “memorizing while reading" paradigm, where the LLM autonomously organizes its memory corpus during a single-pass scan through the documents (Xu et al., 2025; Li et al., 2025; Yu et al., 2025a).”
Toward Personalized LLM-Powered Agents: Foundations, Evaluation, and Future DirectionsPreprint
Xu, Y., Chen, Q., Ma, Z., Liu, D., Wang, W., Wang, X., Xiong, L., & Wang, W. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.22680 · Cited by 35
How Many Instructions Can LLMs Follow at Once?Preprint
Jaroslawicz, D., Whiting, B., Shah, P., & Maamari, K. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2507.11538 · Cited by 28
Adaptive Memory Admission Control for LLM AgentsPreprint
Zhang, G., Jiang, W., Wang, X., Behr, A., Zhao, K., Friedman, J., Chu, X., & Anoun, A. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.04549 · Cited by 29
“In contrast, LLM-native approaches such as A-mem (Xu et al., 2025) and Mem0 (Chhikara et al., 2025) delegate memory admission entirely to language models by generating structured memory attributes.”
Yang, T., Xiao, Y., Bao, Z., Hao, J., & Peng, J. (2025) · Briefings in Bioinformatics · doi:10.1093/bib/bbaf601 · Cited by 33
MemSearcher: Training LLMs to Reason, Search and Manage Memory via End-to-End Reinforcement LearningPreprint
Yuan, Q., Lou, J., Li, Z., Chen, J., Lu, Y., Lin, H., Sun, L., Zhang, D., & Han, X. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2511.02805 · Cited by 35
“This observation highlights the necessity of RL training in equipping models with the ability to effectively interact with both the search engine and memory, thereby enhancing their overall functionality and task-solving ability.”
PlugMem: A Task-Agnostic Plugin Memory Module for LLM AgentsPreprint
Yang, K., Chen, Z., He, X., Jiang, J., Galley, M., Wang, C., Gao, J., Han, J., & Zhai, C. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.03296 · Cited by 35
Survey on Evaluation of LLM-based Agents
Yehudai, A., Eden, L., Li, A., Uziel, G., Zhao, Y., Bar-Haim, R., Cohan, A., & Shmueli-Scheuer, M. (2025) · arXiv · doi:10.48550/arxiv.2503.16416 · Cited by 26
“Recent works, such as ReadAgent (Lee et al., 2024), MemGPT (Packer et al., 2024), and A-MEM (Xu et al., 2025b), investigate these methods and evaluate their efficacy through reasoning and retrieval metrics.”
From threat to trust: assessing security risks of agentic AI systems
Leo, M., Tan, F., Miao, T., & Anand, G. (2026) · International Journal of Information Security · doi:10.1007/s10207-025-01185-y · Cited by 37
“Memory prioritization algorithms dynamically assess which memories are relevant based on task context, exemplified by systems like MemGPT, which utilize cache-like architectures to prioritize recent information [25].”
Jev-Mem: System-One-Controlled Agentic Memory for Efficient AI AgentsPreprint
Jiang, D., Li, Y., & Li, B. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2609.23986 · Cited by 26
“A-MEM (Xu et al., 2025) constructs interconnected memory notes and allows their representations to evolve as new information arrives.”
InfluentialGroupMemBench: Benchmarking LLM Agent Memory in Multi-Party Conversations
Yang, J., Lai, K.-H., Wang, X., Chang, S., Harari, Y., & Gabrilovich, E. (2026) · arXiv · doi:10.48550/arxiv.2605.14498 · Cited by 24
“…working/archival memory with explicit paging, Mem0 [9] extracts atomic facts and applies add/update/delete operations during ingestion, A-Mem [29] links extracted notes into a self-organizing graph, HippoRAG [19] grounds dense retrieval in an entity-centric knowledge graph for multi-hop…”
A Simple Yet Strong Baseline for Long-Term Conversational Memory of LLM AgentsPreprint
Zhou, S. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2511.17208 · Cited by 24
Kwon, T., Choi, D., Kim, H., Kim, S., Moon, S., Kwak, B.-W., Huang, K.-H., & Yeo, J. (2025) · arXiv · doi:10.48550/arxiv.2505.16348
WebLeaper: Empowering Efficiency and Efficacy in WebAgent via Enabling Info-Rich SeekingPreprint
Tao, Z., Shen, H., Li, B., Yin, W., Wu, J., Li, K., Zhang, Z., Yin, H., Ye, R., Zhang, L., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.24697 · Cited by 28
“…Tao et al., 2025; Su et al., 2025; Fang et al., 2025a); (2) advancing agent architecture for improved planning and robustness (Qiao et al., 2025; Xu et al., 2025a; Li et al., 2025a); and (3) developing multi-agent systems for collaborative problem-solving (Wu et al., 2023; Hong et al., 2024).”
Learning Personalized Agents from Human FeedbackPreprint
Liang, K., Kruk, J., Qian, S., Yang, X., Bi, S., Yao, Y., Nie, S., Zhang, M., Liu, L., Fisac, J. F., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.16173 · Cited by 26
“…data—either training models to infer implicit preferences from historical interaction logs (Qiu et al., 2025; Chen et al., 2024; Tan et al., 2024; Li et al., 2024b) or populating agent memory with pre-defined user profiles and past interaction history (Chhikara et al., 2025; Xu et al., 2025b).”
GEMS: Agent-Native Multimodal Generation with Memory and SkillsPreprint
He, Z., Huang, S., Qu, X., Li, Y., Zhu, T., Cheng, Y., & Yang, Y. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.28088 · Cited by 26
Kang, J., Ji, M., Zhao, Z., & Bai, T. (2025) · Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2025.emnlp-main.1318 · Cited by 23
Kang, J., Ji, M., Zhao, Z., & Bai, T. (2025) · arXiv · doi:10.48550/arxiv.2506.06326 · Cited by 23
“This design aims to enable scalable memory extension beyond the fixed context window of LLMs. A-Mem (Agentic Memory) (Xu et al., 2025): it dynamically generates structured notes and links them to form interconnected knowledge networks, enabling continuous memory evolution and adaptive management…”
InfluentialContext Engineering 2.0: The Context of Context EngineeringPreprint
Hua, Q., Ye, L., Fu, D., Xiao, Y., Cai, X., Wu, Y., Lin, J., Wang, J., & Liu, P. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.26493 · Cited by 33
“Systems like MemGPT (Team, 2024) and A-MEM (Chen et al., 2024a) adopt this approach to support indirect communication and coordination between agents.”
SWE Context Bench: A Benchmark for Context Learning in CodingPreprint
Zhu, J., Wu, J., Hu, M., Zhu, S., Pan, J., Shen, W., Yang, Y., Liu, F., Hao, J., Jin, Y., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.08316 · Cited by 23
“In parallel, many recent agent models introduce external context augmentation by retrieving relevant information from outside the current task trajectory (Chhikara et al. [2025], Li et al. [2025], Xu et al. [2025]).”
PersonaX: A Recommendation Agent-Oriented User Modeling Framework for Long Behavior Sequence
Shi, Y., Xu, W., Zeqi, Z., Zi, X., Wu, Q., & Xu, M. (2025) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2025.findings-acl.300 · Cited by 51
Self-citationPreference-Aware Memory Update for Long-Term LLM Agents
Sun, H., Zhang, Z., & Zeng, S. (2026) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2026.findings-acl.38 · Cited by 30
Beyond Dialogue Time: Temporal Semantic Memory for Personalized LLM Agents
Su, M., Guo, Y., Hou, Z., Bai, L., Li, Z., Zhang, Y., Yin, G., Lin, W., Jin, X., Guo, J., et al. (2026) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2026.findings-acl.1496 · Cited by 28
Self-citationGAM: Hierarchical Graph-based Agentic Memory for LLM Agents
Wu, Z., Zhang, H., Lin, F., Xu, W., Xinran, X., Chen, Y., Zou, H. P., Chen, S., Zhang, W., Liu, X., et al. (2026) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2026.acl-long.1600 · Cited by 37
Self-citationAgentOCR: Reimagining Agent History via Optical Self-Compression
Feng, L., Yang, F., Chen, F., Cheng, X., Xu, H., Wan, Z., Yan, M., & An, B. (2026) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2026.acl-long.230 · Cited by 27
MemLens: Benchmarking Multimodal Long-Term Memory in Large Vision-Language ModelsPreprint
Ren, X., Wang, Z., Du, Y., Xie, Z., Liu, C., Yang, X., Feng, H., Pan, W., Zheng, T., Xu, B., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.14906 · Cited by 23
Governed Shared Memory for Multi-Agent LLM SystemsPreprint
Margalit, Y., Cohen-Inger, N., Avram, E., Taig, R., & Margalit, O. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2606.24535 · Cited by 23
Diefenbacher, S., Hallin, A., Kasieczka, G., Krämer, M., Lauscher, A., & Lukas, T. (2026) · SciPost Physics · doi:10.21468/scipostphys.20.3.091 · Cited by 22
HiMem: Hierarchical Long-Term Memory for LLM Long-Horizon AgentsPreprint
Zhang, N., Yang, X., Tan, Z., Deng, W., & Wang, W. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2601.06377 · Cited by 22
“A-MEM [30] augments event-level memory with entities, relations, and temporal features to support time-aware retrieval and reasoning.”
According to Me: Long-Term Personalized Referential Memory QAPreprint
Mei, J., Chen, J., Yang, G., Hou, X., Li, M., & Byrne, B. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.01990 · Cited by 24
“A-Mem (Xu et al., 2025), Mem0 (Chhikara et al., 2025), and HippoRAG2 (Guti´errez et al., 2025) employ Linked Memory Organizations, whereas Self-RAG (Asai et al., 2024) adopts a reflective answer generation mechanism.”
MEMENTO: Teaching LLMs to Manage Their Own ContextPreprint
Kontonis, V., Zeng, Y., Garg, S., Chen, L., Tang, H., Wang, Z., Awadallah, A. H., Horvitz, E., Langford, J. C., & Papailiopoulos, D. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2604.09852 · Cited by 21
“Context management for long-running models is typically handled through external infrastructure: separate summarizers, memory modules, or orchestration logic (Wu et al., 2025; Xu et al., 2025).”
AlpsBench: An LLM Personalization Benchmark for Real-Dialogue Memorization and Preference Alignment
Xiao, J., Yu, X., Wang, C., Zheng, W., Lin, X., Liu, K., Ding, H., Zhang, Y., Wang, W., Feng, F., et al. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3808634 · Cited by 23
“For memory-oriented systems , we implemented MemoryOS [15], EverMemOS [10], A-Mem [50], Light-Mem [7], MemOS [19], Mem0 [5], and Mem0g [5].”
Self-citationChoosing How to Remember: Adaptive Memory Structures for LLM AgentsPreprint
Lu, M., Wu, M., Liu, F., Xu, J., Li, W., Wang, H., Hu, Z., Ding, Y., Sun, Y., Lu, J., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.14038 · Cited by 20
“A-MEM supports hierarchical and relational reasoning by maintaining structured connections between notes, facilitating effective recall during interactions.”
InfluentialSelf-citationMulti-User Large Language Model AgentsPreprint
Yang, S., Zhu, S., Zhu, H., Enríquez, J. R., Wang, D., Pentland, A., Bakker, M. A., & Pei, J. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2604.08567 · Cited by 21
“Large language Model (LLM) and LLM-based agent systems, equipped with strong abilities in planning (Huang et al., 2024), decision-making (Yang et al., 2023), tool use (Qin et al., 2024), and memory management (Xu et al., 2025), are increasingly capable of navigating complex and dynamic environments.”
Dynamic Skill Lifecycle Management for Agentic Reinforcement LearningPreprint
Shen, J., Zhang, T., Zhao, X., & Cheng, H. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.10923 · Cited by 23
Lumer, E., Gulati, A., Subbiah, V. K., Basavaraju, P. H., & Burke, J. A. (2025) · arXiv · doi:10.48550/arxiv.2507.21428 · Cited by 19
“Furthermore, some methods of updating or adding memories for tool personalization include an LLM agent calling a tool to update memories (Letta, 2025; Xu et al., 2025).”
When Continual Learning Moves to Memory: A Study of Experience Reuse in LLM Agents
Hu, Q., Long, Q., & Wang, W. (2026) · arXiv · doi:10.48550/arxiv.2604.27003 · Cited by 19
“…with external memory, from retrieval-augmented language models (Borgeaud et al., 2022; Wang et al., 2023b) to agent systems that store and reuse experience across interactions (Zhong et al., 2024; Shinn et al., 2023; Wang et al., 2023a; Packer et al., 2023; Zhao et al., 2024; Xu et al., 2025).”
Hu, B. A., Liu, Y., & Rong, H. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.09757 · Cited by 22
“Their identity and behavior emerge from three elements: fixed profile definition prompts, memory constructs [90], and interactions with their environment.”
AI Meets Brain: Memory Systems from Cognitive Neuroscience to Autonomous Agents
Liang, J., Li, H., Li, C., Zhou, J., Jiang, S., Wang, Z., Ji, C., Zhu, Z., Liu, R., Ren, T., et al. (2025) · arXiv · doi:10.48550/arxiv.2512.23343
Position: Agentic Evolution is the Path to Evolving LLMsPreprint
Lin, M., Lu, H., Shi, Z., He, B., Mao, R., Zhang, Z., Wu, Z., Tang, X., Liu, H., Dai, Z., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.00359 · Cited by 26
Trajectory-Informed Memory Generation for Self-Improving Agent SystemsPreprint
Fang, G., Isahagian, V., Jayaram, K. R., Kumar, R., Muthusamy, V., Oum, P., & Thomas, G. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.10600 · Cited by 19
“Generic memory systems [2, 15] store facts extracted from conversations in vector databases for later retrieval, but these systems lack several critical capabilities: they have no understanding of agent execution patterns and reasoning flows, they cannot perform causal analysis to identify which…”
Metadata
- Authors
- Xu, W., Liang, Z., Mei, K., Gao, H., Tan, J., & Zhang, Y.
- Affiliations
- Rutgers, The State University of New Jersey (US)
- Ant Group (China) (CN)
- Salesforce (United States) (US)
- Venue
- Advances in Neural Information Processing Systems 38
- Published
- 2025
- Pages
- 20004-20031
- ISSN
- 1049-5258 (ISSN-L)
- OpenAlex
- W4407759090
- Cited by
- 51
- References
- 0
- Citations by year
- 2026: 48 · 2025: 3
- FWCI
- 48.02
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- neural information processing systems (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › AI-based Problem Solving and Planning
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
- Keywords
- agentic memory, LLM agents, knowledge networks, memory organization, dynamic indexing, evolution of memory, knowledge representation, memory management
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-09-28, under version 3 of the claim definition (prompt extract-v3).
A second model labelled each claim on 2026-09-28 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.