[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-1846":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":10,"languages":10,"totalLinesOfCode":10,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":15,"stars7d":16,"stars30d":17,"stars90d":14,"forks30d":14,"starsTrendScore":18,"compositeScore":19,"rankGlobal":10,"rankLanguage":10,"license":10,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":22,"hasPages":22,"topics":23,"createdAt":10,"pushedAt":10,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":14,"starSnapshotCount":14,"syncStatus":27,"lastSyncTime":28,"discoverSource":29},1846,"Awesome-Agent-Memory-Papers","yyyujintang\u002FAwesome-Agent-Memory-Papers","yyyujintang","Awesome Papers related to Agent Memory: methods, benchmarks and surveys. Website: https:\u002F\u002Fyyyujintang.github.io\u002FAwesome-Agent-Memory-Papers\u002F","",null,202,6,1,0,5,9,25,15,2.54,false,"main",true,[],"2026-06-12 02:00:33","# Awesome Agent Memory Papers\n\n[![Stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fyyyujintang\u002FAwesome-Agent-Memory-Papers?style=social)](https:\u002F\u002Fgithub.com\u002Fyyyujintang\u002FAwesome-Agent-Memory-Papers\u002Fstargazers) ![Last updated](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flast%20updated-2026--04--21-blue) ![Papers](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpapers-90-success)\n\nA curated list of papers on **memory for LLM \u002F multimodal agents** — methods, benchmarks, and surveys — covering episodic, semantic, procedural, and multimodal memory, with both parametric (internal) and retrieval-based (external) storage, learned via prompting, supervised finetuning, or reinforcement learning.\n\n**90 papers** · 7 surveys · 31 benchmarks · 52 methods · last updated **2026-04-21**\n\nInteractive dashboard with multi-tag filtering: **\u003Chttps:\u002F\u002Fyyyujintang.github.io\u002FAwesome-Agent-Memory-Papers\u002F>**\n\n\n> Contributions welcome — open an issue or PR with new papers.\n\n## Contents\n\n- [Surveys](#surveys)\n- [Benchmarks](#benchmarks)\n  - [QA-based Memory Evaluation](#qa-based-memory-evaluation) (5)\n  - [Web Navigation](#web-navigation) (7)\n  - [Desktop \u002F Mobile GUI](#desktop-mobile-gui) (6)\n  - [Embodied & Game Environments](#embodied-game-environments) (6)\n  - [General Long-Horizon \u002F Office](#general-long-horizon-office) (7)\n- [Methods](#methods)\n  - [Multimodal Memory](#multimodal-memory) (16)\n  - [Procedural Memory](#procedural-memory) (10)\n  - [Episodic Memory](#episodic-memory) (18)\n  - [Semantic Memory](#semantic-memory) (2)\n  - [Internal \u002F Parametric Memory](#internal-parametric-memory) (4)\n  - [Other Methods](#other-methods) (2)\n- [Tag Legend](#tag-legend)\n\n## Surveys\n\n- [Rethinking Memory Mechanisms of Foundation Agents in the Second Half](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.06052)  \n  *2026-01-14* · Jiawei Han, Philip Yu  \n  `Survey`\n- [AI Meets Brain: Memory Systems from Cognitive Neuroscience to Autonomous Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.23343)  \n  *2025-12-29* · [[code]](https:\u002F\u002Fgithub.com\u002FAgentMemory\u002FHuaman-Agent-Memory)  \n  `Survey`\n- [Memory in the Age of AI Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.13564)  \n  *2025-12-15* · Shuicheng Yan, Guibin Zhang · [[code]](https:\u002F\u002Fgithub.com\u002FShichun-Liu\u002FAgent-Memory-Paper-List)  \n  `Survey`\n- [Measuring Agents in Production](https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.04123)  \n  *2025-12-02* · Shuicheng Yan, Guibin Zhang  \n  `Survey`\n- [Retrieval Augmented Generation and Understanding in Vision: A Survey and New Outlook](https:\u002F\u002Farxiv.org\u002Fabs\u002F2503.18016)  \n  *2025-03-23* · Xuming Hu  \n  `Survey`\n- [Episodic memory in AI agents poses risks that should be studied and mitigate](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.11739)  \n  *2025-01-20*  \n  `Survey`\n- [A Survey on the Memory Mechanism of Large Language Model based Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2404.13501)  \n  *2024-04-21*  \n  `Survey`\n\n## Benchmarks\n\nEvaluation suites for agent memory, split by interaction mode.\n\n### QA-based Memory Evaluation\n\n- [KnowMe-Bench: Benchmarking Person Understanding for Lifelong Digital Companions](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.04745)  \n  *2026-01-08*  \n  `Benchmark` `QA`\n- [Mem-Gallery: Benchmarking Multimodal Long-Term Conversational Memory for MLLM Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.03515)  \n  *2026-01-07*  \n  `Benchmark` `QA`\n- [Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.05257)  \n  *2025-07-07* · **ICLR26** · Yuanzhe Hu · [[code]](https:\u002F\u002Fgithub.com\u002FHUST-AI-HYZ\u002FMemoryAgentBench)  \n  `Benchmark` `QA`\n- [LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2410.10813)  \n  *2024-10-14*  \n  `Benchmark` `QA`\n- [(LoCoMo) Evaluating Very Long-Term Conversational Memory of LLM Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2402.17753)  \n  *2024-02-27*  \n  `Benchmark` `QA`\n\n### Web Navigation\n\n- WebChoreArena: Evaluating Web Browsing Agents on Realistic Tedious Web Tasks  \n  *2025-06-02*  \n  `Benchmark` `Web`\n- [RealWebAssist: A Benchmark for Long-Horizon Web Assistance with Real-World Users](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.10445)  \n  *2025-04-14*  \n  `Benchmark` `Web`\n- [The BrowserGym Ecosystem for Web Agent Research](https:\u002F\u002Farxiv.org\u002Fabs\u002F2412.05467)  \n  *2024-12-06*  \n  `Benchmark` `Web`\n- [VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web Tasks](https:\u002F\u002Farxiv.org\u002Fabs\u002F2401.13649)  \n  *2024-01-24* · **ACL24** · [[code]](https:\u002F\u002Fgithub.com\u002Fweb-arena-x\u002Fvisualwebarena)  \n  `Benchmark` `Web`\n- [WebArena: A Realistic Web Environment for Building Autonomous Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2307.13854)  \n  *2023-07-25* · **ICLR24** · Shuyan Zhou, Duke  \n  `Benchmark` `Web`\n- [Mind2Web: Towards a Generalist Agent for the Web](https:\u002F\u002Farxiv.org\u002Fabs\u002F2306.06070)  \n  *2023-06-09* · **NeurIPS23, Spotlight**  \n  `Benchmark` `Web`\n- [WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2207.01206)  \n  *2022-07-04* · **NeurIPS22** · [[code]](https:\u002F\u002Fgithub.com\u002Fprinceton-nlp\u002FWebShop)  \n  `Benchmark` `Web`\n\n### Desktop \u002F Mobile GUI\n\n- [Gym-Anything: Turn any Software into an Agent Environment](https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.06126)  \n  *2026-04-07*  \n  `Benchmark` `GUI`\n- [MemGUI-Bench: Benchmarking Memory of Mobile GUI Agents in Dynamic Environments](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.06075)  \n  *2026-02-03* · [[code]](https:\u002F\u002Fgithub.com\u002Flgy0404\u002FMemGUI-Bench)  \n  `Benchmark` `GUI`\n- [OS-Marathon: Benchmarking Computer-Use Agents on Long-Horizon Repetitive Tasks](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.20650)  \n  *2026-01-28*  \n  `Benchmark` `GUI`\n- [LongHorizonUI: A Unified Framework for Robust long-horizon Task Automation of GUI Agent](https:\u002F\u002Fopenreview.net\u002Fforum?id=BK7Mk5d4WE)  \n  *2026-01-26* · **ICLR26**  \n  `Benchmark` `GUI`\n- [VisualAgentBench: Towards Large Multimodal Models as Visual Foundation Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2408.06327)  \n  *2024-08-12* · THU-Jie Tang · [[code]](https:\u002F\u002Fgithub.com\u002FTHUDM\u002FVisualAgentBench)  \n  `Benchmark` `GUI`\n- [OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments](https:\u002F\u002Farxiv.org\u002Fabs\u002F2404.07972)  \n  *2024-04-11* · **NeurIPS24** · [[code]](https:\u002F\u002Fgithub.com\u002Fxlang-ai\u002FOSWorld)  \n  `Benchmark` `GUI`\n\n### Embodied & Game Environments\n\n- [AGENTVISTA: Evaluating Multimodal Agents in Ultra-Challenging\r Realistic Visual Scenarios](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.23166)  \n  *2026-02-26* · Junxian He, May Fung  \n  `Benchmark` `Embodied`\n- [MentisOculi: Revealing the Limits of Reasoning with Mental Imagery](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.02465)  \n  *2026-02-02*  \n  `Benchmark` `Embodied`\n- [CAPTURe: Evaluating Spatial Reasoning in Vision Language Models via Occluded Object Counting](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.15485)  \n  *2025-04-21* · **ICCV25** · Mohit Bansal  \n  `Benchmark` `Embodied`\n- [ALFWorld: Aligning Text and Embodied Environments for Interactive Learning](https:\u002F\u002Farxiv.org\u002Fabs\u002F2010.03768)  \n  *2020-10-08* · **ICLR21**  \n  `Benchmark` `Embodied`\n- [ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Tasks](https:\u002F\u002Farxiv.org\u002Fabs\u002F1912.01734)  \n  *2019-12-03* · **CVPR20**  \n  `Benchmark` `Embodied`\n- [TextWorld: A Learning Environment for Text-based Game](https:\u002F\u002Farxiv.org\u002Fabs\u002F1806.11532)  \n  *2018-06-29* · **IJCAI18**  \n  `Benchmark` `Embodied`\n\n### General Long-Horizon \u002F Office\n\n- [AMA-Bench: Evaluating Long-Horizon Memory for Agentic Applications](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.22769)  \n  *2026-02-26*  \n  `Benchmark` `Long-Horizon`\n- [MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic Tasks](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.16313)  \n  *2026-02-12* · Yu Wang, Yuanzhe Hu  \n  `Benchmark` `Long-Horizon`\n- [A Framework for Studying AI Agent Behavior: Evidence from Consumer Choice Experiments](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.25609)  \n  *2025-09-30* · **ICLR26** · Nikhil, ABxLab  \n  `Benchmark` `Long-Horizon`\n- [OdysseyBench: Evaluating LLM Agents on Long-Horizon Complex Office Application Workflows](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.09124)  \n  *2025-08-12*  \n  `Benchmark` `Long-Horizon`\n- [TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks](https:\u002F\u002Farxiv.org\u002Fabs\u002F2412.14161)  \n  *2024-12-18*  \n  `Benchmark` `Long-Horizon`\n- [MemSim: A Bayesian Simulator for Evaluating Memory of LLM-based Personal Assistants](https:\u002F\u002Farxiv.org\u002Fabs\u002F2409.20163)  \n  *2024-09-30* · **NeurIPS25** · [[code]](https:\u002F\u002Fgithub.com\u002Fnuster1128\u002FMemSim)  \n  `Benchmark` `Long-Horizon`\n- [AgentBench: Evaluating LLMs as Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2308.03688)  \n  *2023-08-07* · **ICLR24** · [[code]](https:\u002F\u002Fgithub.com\u002FTHUDM\u002FAgentBench)  \n  `Benchmark` `Long-Horizon`\n\n## Methods\n\nEach paper is placed in exactly **one** primary section (Multimodal > Procedural > Episodic > Semantic > External > Internal). Tag badges on each entry show the full tag vector — use the website for true multi-axis filtering.\n\n### Multimodal Memory\n\n- [Omni-SimpleMem: Autoresearch-Guided Discovery of Lifelong Multimodal Agent Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.01007)  \n  *2026-04-01*  \n  `Method` `External` `Prompt-based` `Episodic` `Multimodal` `Procedural` `Semantic`\n- [Visual Generation Unlocks Human-Like Reasoning through Multimodal World Models](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.19834)  \n  *2026-01-27* · Mingsheng Long  Bytedance Seed  \n  `Method` `Internal` `SFT` `Multimodal`\n- [MemOCR: Layout-Aware Visual Memory for Efficient Long-Horizon Reasoning](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.21468)  \n  *2026-01-26*  \n  `Method` `External` `Prompt-based` `Episodic` `Multimodal`\n- [MemVerse: Multimodal Memory for Lifelong Learning Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.03627)  \n  *2025-12-03* · [[code]](https:\u002F\u002Fgithub.com\u002FKnowledgeXLab\u002FMemVerse)  \n  `Method` `External` `Prompt-based` `Episodic` `Multimodal` `Procedural` `Semantic`\n- [ViLoMem: Agentic Learner with Grow-and-Refine Multimodal Semantic Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.21678)  \n  *2025-11-26* · **CVPR26** · [[code]](https:\u002F\u002Fgithub.com\u002Fweihao-bo\u002FViLoMem\u002Ftree\u002Fmain)  \n  `Method` `External` `Multimodal` `Semantic`\n- [LongVT: Incentivizing \"Thinking with Long Videos\" via Native Tool Calling](https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.20785)  \n  *2025-11-25* · **CVPR26**  \n  `Method` `External` `Prompt-based` `Multimodal` `Procedural`\n- [VisMem: Latent Vision Memory Unlocks Potential of Vision-Language Models](https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.11007)  \n  *2025-11-14* · Shuicheng Yan  \n  `Method` `Internal` `SFT` `Multimodal`\n- [VAGEN: Reinforcing World Model Reasoning for Multi-Turn VLM Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.16907)  \n  *2025-10-19* · **NeurIPS25** · [[code]](https:\u002F\u002Fgithub.com\u002Fmll-lab-nu\u002FVAGEN)  \n  `Method` `Internal` `RL-based` `Episodic` `Multimodal`\n- [VideoLucy: Deep Memory Backtracking for Long Video Understanding](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.12422)  \n  *2025-10-14* · **NeurIPS25**  \n  `Method` `External` `SFT` `Episodic` `Multimodal`\n- [(M3-Agent) Seeing, Listening, Remembering, and Reasoning: A Multimodal Agent with Long-Term Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.09736)  \n  *2025-08-13* · **ICLR26** · ByteDance Seed · [[code]](https:\u002F\u002Fgithub.com\u002Fbytedance-seed\u002Fm3-agent)  \n  `Method` `External` `SFT` `Episodic` `Multimodal` `Semantic`\n- [MAViS: A Multi-Agent Framework for Long-Sequence Video Storytelling](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.08487)  \n  *2025-08-11*  \n  `Method` `External` `Prompt-based` `Episodic` `Multimodal`\n- [Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Token](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.17218)  \n  *2025-06-20* · Chuang Gan · [[code]](https:\u002F\u002Fgithub.com\u002FUMass-Embodied-AGI\u002FMirage)  \n  `Method` `Internal` `Multimodal`\n- [3DLLM-Mem: Long-Term Spatial-Temporal Memory for Embodied 3D Large Language Model](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.22657)  \n  *2025-05-28*  \n  `Method` `External` `Prompt-based` `Episodic` `Multimodal` `Semantic`\n- [Towards General Continuous Memory for Vision-Language Models](https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.17670)  \n  *2025-05-23* · **NeurIPS25**  \n  `Method` `External` `Internal` `SFT` `Episodic` `Multimodal` `Semantic`\n- [SAM2Act: Integrating Visual Foundation Model with A Memory Architecture for Robotic Manipulation](https:\u002F\u002Farxiv.org\u002Fabs\u002F2501.18564)  \n  *2025-01-30* · [[code]](https:\u002F\u002Fgithub.com\u002Fsam2act\u002Fsam2act)  \n  `Method` `External` `Prompt-based` `Episodic` `Multimodal`\n- [Optimus-1: Hybrid Multimodal Memory Empowered Agents Excel in Long-Horizon Tasks](https:\u002F\u002Farxiv.org\u002Fabs\u002F2408.03615)  \n  *2024-08-07* · **NeurIPS24** · [[code]](https:\u002F\u002Fgithub.com\u002FJiuTian-VL\u002FOptimus-1)  \n  `Method` `External` `Prompt-based` `Multimodal`\n\n### Procedural Memory\n\n- [A Subgoal-driven Framework for Improving Long-Horizon LLM Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2603.19685)  \n  *2026-03-20*  \n  `Method` `External` `Prompt-based` `Training-free` `Episodic` `Procedural`\n- [Plan-MCTS: Plan Exploration for Action Exploitation in Web Navigation](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.14083)  \n  *2026-02-15* · Weinan Zhang  \n  `Method` `RL-based` `Procedural`\n- [MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.02474)  \n  *2026-02-02*  \n  `Method` `External` `RL-based` `Procedural`\n- [TokMem: Tokenized Procedural Memory for Large Language Models](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.00444)  \n  *2025-10-01*  \n  `Method` `Internal` `SFT` `Procedural`\n- [ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.25140)  \n  *2025-09-29* · **ICLR26** · Siru Ouyang  \n  `Method` `External` `Prompt-based` `Episodic` `Procedural`\n- [Memory Management and Contextual Consistency for Long-Running Low-Code Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.25250)  \n  *2025-09-27*  \n  `Method` `External` `Prompt-based` `Episodic` `Procedural`\n- [Memory OS of AI Agent](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.06326)  \n  *2025-05-30* · **EMNLIP25 Main**  \n  `Method` `External` `Prompt-based` `Episodic` `Procedural` `Semantic`\n- [A-MEM: Agentic Memory for LLM Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2502.12110)  \n  *2025-02-17* · **NeurIPS25** · [[code]](https:\u002F\u002Fgithub.com\u002FWujiangXu\u002FA-mem)  \n  `Method` `External` `Prompt-based` `Episodic` `Procedural` `Semantic`\n- [Agent Workflow Memory (AWM)](https:\u002F\u002Farxiv.org\u002Fabs\u002F2409.07429)  \n  *2024-09-11* · **ICML26** · [[code]](https:\u002F\u002Fgithub.com\u002Fzorazrw\u002Fagent-workflow-memory)  \n  `Method` `External` `Prompt-based` `Procedural`\n- [Synapse: Trajectory-as-Exemplar Prompting with Memory for Computer Control](https:\u002F\u002Farxiv.org\u002Fabs\u002F2306.07863)  \n  *2023-06-13* · [[code]](https:\u002F\u002Fgithub.com\u002Fltzheng\u002FSynapse)  \n  `Method` `External` `Prompt-based` `Episodic` `Procedural`\n\n### Episodic Memory\n\n- [Gated Memory Policy](https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.18933)  \n  *2026-04-21* · Shuran Song  \n  `Method` `Internal` `RL-based` `Episodic`\n- [HiGMem: A Hierarchical and LLM-Guided Memory System for Long-Term Conversational Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.18349)  \n  *2026-04-20*  \n  `Method` `External` `Prompt-based` `Training-free` `Episodic` `Semantic`\n- [PlugMem: A Task-Agnostic Plugin Memory Module for LLM Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2603.03296)  \n  *2026-02-23* · [[code]](https:\u002F\u002Fgithub.com\u002FTIMAN-group\u002FPlugMem)  \n  `Method` `External` `Prompt-based` `Training-free` `Episodic` `Semantic`\n- [Modeling Distinct Human Interaction in Web Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.17588)  \n  *2026-02-19*  \n  `Method` `External` `Prompt-based` `Episodic`\n- [REMem: Reasoning with Episodic Memory in Language Agent](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.13530)  \n  *2026-02-13* · Yu Su, Huan Sun  \n  `Method` `External` `Prompt-based` `Episodic`\n- [TraceMem: Weaving Narrative Memory Schemata from User Conversational Traces](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.09712)  \n  *2026-02-10* · HKU  \n  `Method` `External` `Prompt-based` `Episodic` `Semantic`\n- [Learning to Continually Learn via Meta-learning Agentic Memory Designs](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.07755)  \n  *2026-02-08* · [[code]](https:\u002F\u002Fgithub.com\u002Fzksha\u002Falma)  \n  `Method` `External` `RL-based` `Episodic`\n- [Dep-Search: Learning Dependency-Aware Reasoning Traces with Persistent Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.18771)  \n  *2026-01-27*  \n  `Method` `External` `Prompt-based` `Episodic`\n- CAST: Character-and-Scene Episodic Memory for Agents  \n  *2026-01-14*  \n  `Method` `External` `Prompt-based` `Episodic`\n- [SimpleMem: Efficient Lifelong Memory for LLM Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.02553)  \n  *2026-01-05*  \n  `Method` `External` `Prompt-based` `Episodic` `Semantic`\n- [Hindsight is 20\u002F20: Building Agent Memory that Retains, Recalls, and Reflects](https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.12818v1)  \n  *2025-12-14*  \n  `Method` `External` `Prompt-based` `Episodic` `Semantic`\n- [A neural network model of free recall learns multiple memory strategies](https:\u002F\u002Fwww.biorxiv.org\u002Fcontent\u002F10.1101\u002F2025.09.25.678592v1.abstract)  \n  *2025-09-25* · [[code]](https:\u002F\u002Fgithub.com\u002FVeritaria\u002Frnn-free-recall)  \n  `Method` `Internal` `Episodic`\n- [PRIME: Large Language Model Personalization with Cognitive Dual-Memory and Personalized Thought Process](https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.04607)  \n  *2025-07-07* · **EMNLP25, Main**  \n  `Method` `External` `Prompt-based` `Episodic` `Semantic`\n- [Ella: Embodied Social Agents with Lifelong Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.24019)  \n  *2025-06-30* · Chuang Gan  \n  `Method` `External` `Prompt-based` `Episodic` `Semantic`\n- [Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2504.19413)  \n  *2025-04-28*  \n  `Method` `External` `Prompt-based` `Episodic` `Semantic`\n- [R3Mem: Bridging Memory Retention and Retrieval via Reversible Compressio\r n](https:\u002F\u002Farxiv.org\u002Fabs\u002F2502.15957)  \n  *2025-02-21*  \n  `Method` `External` `Prompt-based` `Episodic`\n- [HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models](https:\u002F\u002Farxiv.org\u002Fabs\u002F2405.14831)  \n  *2024-05-23* · **NeurIPS24** · Yu Su · [[code]](https:\u002F\u002Fgithub.com\u002FOSU-NLP-Group\u002FHippoRAG)  \n  `Method` `External` `Prompt-based` `Training-free` `Episodic` `Semantic`\n- [MemoryBank: Enhancing Large Language Models with Long-Term Memory](https:\u002F\u002Farxiv.org\u002Fabs\u002F2305.10250)  \n  *2023-05-17*  \n  `Method` `External` `Prompt-based` `Episodic` `Semantic`\n\n### Semantic Memory\n\n- [Explicit v.s. Implicit Memory: Exploring Multi-hop Complex Reasoning Over Personalized Information](https:\u002F\u002Farxiv.org\u002Fabs\u002F2508.13250)  \n  *2025-08-15* · **SIGKDD 26** · Zeyu Zhang  \n  `Method` `External` `Internal` `Prompt-based` `Semantic`\n- [From RAG to Memory: Non-Parametric Continual Learning for Large Language Models  (HippoRAG 2)](https:\u002F\u002Farxiv.org\u002Fabs\u002F2502.14802)  \n  *2025-02-20* · **ICML25** · [[code]](https:\u002F\u002Fgithub.com\u002FOSU-NLP-Group\u002FHippoRAG)  \n  `Method` `External` `Internal` `Prompt-based` `Semantic`\n\n### Internal \u002F Parametric Memory\n\n- [When to Memorize and When to Stop: Gated Recurrent Memory for Long-Context Reasoning](https:\u002F\u002Farxiv.org\u002Fabs\u002F2602.10560)  \n  *2026-02-11* · Bytedance Seed  \n  `Method` `Internal` `SFT`\n- [QwenLong-L1.5: Post-Training Recipe for Long-Context Reasoning and Memory Management](https:\u002F\u002Farxiv.org\u002Fabs\u002F2512.12967v1)  \n  *2025-12-25*  \n  `Method` `Internal` `SFT`\n- [MemGen: Weaving Generative Latent Memory for Self-Evolving Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2509.24704)  \n  *2025-09-29* · Shuicheng Yan, Guibin Zhang · [[code]](https:\u002F\u002Fgithub.com\u002FKANABOON1\u002FMemGen)  \n  `Method` `Internal`\n- [Scaling Test-time Compute for LLM Agents](https:\u002F\u002Farxiv.org\u002Fabs\u002F2506.12928)  \n  *2025-06-15* · **ICLR26**  \n  `Method` `Internal` `Prompt-based`\n\n### Other Methods\n\n- [Agentic Reasoning for Large Language Models](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.12538)  \n  *2026-01-18* · Heng Ji  \n  `Method` `Prompt-based` `Training-free`\n- [AgentRL: Scaling Agentic Reinforcement Learning with a Multi-Turn, Multi-Task Framework](https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.04206)  \n  *2025-10-05* · Jie Tang  \n  `Method` `RL-based`\n\n## Tag Legend\n\n| Axis | Values |\n|---|---|\n| **Category** | `Survey` · `Benchmark` · `Method` |\n| **Benchmark Type** | `QA` · `Web` · `GUI` · `Embodied` · `Long-Horizon` |\n| **Storage** | `Internal` (parametric — weights \u002F latent tokens) · `External` (non-parametric — retrieval) |\n| **Learning** | `Prompt-based` · `RL-based` · `SFT` · `Training-free` |\n| **Memory Type** | `Episodic` · `Semantic` · `Procedural` · `Multimodal` |\n\n## Citation\n\nIf this list is useful in your work, please consider starring the repo.\n\n","该项目是一个关于代理记忆相关论文的精选列表，涵盖了方法、基准测试和综述。它包括90篇论文，涉及情景记忆、语义记忆、程序记忆以及多模态记忆，并且这些记忆可以通过提示学习、监督微调或强化学习获得。项目还提供了一个交互式仪表板，支持多标签过滤，方便研究人员根据需求筛选相关内容。适合于对大语言模型或跨模态代理的记忆机制感兴趣的科研人员及开发者使用。",2,"2026-06-11 02:46:24","CREATED_QUERY"]