mlsanityData to 7 Oct 2026

MemGen: Weaving Generative Latent Memory for Self-Evolving Agents

arXiv 2509.24704 ↗First posted 29 Sep 2025cs.CLcited by 99

On the map: LLM agents and tool use → Agent memory systems

also in 1 more
Abstract (arXiv)

Agent memory shapes how Large Language Model (LLM)-powered agents, akin to the human brain, progressively refine themselves through environment interactions. Existing paradigms remain constrained: parametric memory forcibly adjusts model parameters, and retrieval-based memory externalizes experience into structured databases, yet neither captures the fluid interweaving of reasoning and memory that underlies human cognition. To address this gap, we propose MemGen, a dynamic generative memory framework that equips agents with a human-esque cognitive faculty. It consists of a textitmemory trigger, which monitors the agent's reasoning state to decide explicit memory invocation, and a textitmemory weaver, which takes the agent's current state as stimulus to construct a latent token sequence as machine-native memory to enrich its reasoning. In this way, MemGen enables agents to recall and augment latent memory throughout reasoning, producing a tightly interwoven cycle of memory and cognition. Extensive experiments across eight benchmarks show that MemGen surpasses leading external memory systems such as ExpeL and AWM by up to 38.22%, exceeds GRPO by up to 13.44%, and exhibits strong cross-domain generalization ability. More importantly, we find that without explicit supervision, MemGen spontaneously evolves distinct human-like memory faculties, including planning memory, procedural memory, and working memory, suggesting an emergent trajectory toward more naturalistic forms of machine cognition.

Key facts. Cited by 99 papers in the map (data through 7 Oct 2026). Placed in Agent memory systems, one of 16 directions in LLM agents and tool use. Builds on 60 of its 78 references that are in the map. First posted on arXiv 29 Sep 2025.

Builds on

60 of 78 references.

Closest ancestors · 6

Same direction, newest first

  1. 2025Memp: Exploring Agent Procedural Memory8 Aug
  2. 2025MEM1: Learning to Synergize Memory and Reasoning for Efficient Long-Horizon Agents18 Jun
  3. 2025G-Memory: Tracing Hierarchical Memory for Multi-Agent Systems9 Jun
  4. 2025How Memory Management Impacts LLM Agents: An Empirical Study of Experience-Following Behavior21 May
  5. 2025A-MEM: Agentic Memory for LLM Agents17 Feb
  6. 2024A Survey on the Memory Mechanism of Large Language Model based Agents21 Apr

Other references · 54

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,146
  4. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  5. 2024Qwen2.5 Technical Reportcited by 4,713
  6. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,895
  7. 2024LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Codecited by 1,906
  8. 2024SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineeringcited by 1,628
  9. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,362
  10. 2020ALFWorld: Aligning Text and Embodied Environments for Interactive Learningcited by 1,143
  11. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memoriescited by 1,103
  12. 2024Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Headscited by 821
38 more
  1. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 763
  2. 2017Program Induction by Rationale Generation : Learning to Solve and Explain Algebraic Word Problemscited by 642
  3. 2024EAGLE: Speculative Sampling Requires Rethinking Feature Uncertaintycited by 564
  4. 2023MemoryBank: Enhancing Large Language Models with Long-Term Memorycited by 539
  5. 2024Break the Sequential Dependency of LLM Inference Using Lookahead Decodingcited by 295
  6. 2023Prompt Cache: Modular Attention Reuse for Low-Latency Inferencecited by 255
  7. 2025MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agentcited by 235
  8. 2025CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillationcited by 216
  9. 2025SkillWeaver: Web Agents can Self-Improve by Discovering and Honing Skillscited by 174
  10. 2025A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligencecited by 155
  11. 2025Mobile-Agent-E: Self-Evolving Mobile Assistant for Complex Taskscited by 128
  12. 2025From Human Memory to AI Memory: A Survey on Memory Mechanisms in the Era of LLMscited by 119
  13. 2025KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Codingcited by 115
  14. 2025Alita: Generalist Agent Enabling Scalable Agentic Reasoning with Minimal Predefinition and Maximal Self-Evolutioncited by 114
  15. 2025SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMscited by 110
  16. 2025Agentic Reasoning and Tool Integration for LLMs via Reinforcement Learningcited by 107
  17. 2024Agent-FLAN: Designing Data and Methods of Effective Agent Tuning for Large Language Modelscited by 97
  18. 2024RAGCache: Efficient Knowledge Caching for Retrieval-Augmented Generationcited by 97
  19. 2025Towards Scientific Intelligence: A Survey of LLM-based Scientific Agentscited by 96
  20. 2024WISE: Rethinking the Knowledge Memory for Lifelong Model Editing of Large Language Modelscited by 90
  21. 2024MEMORYLLM: Towards Self-Updatable Large Language Modelscited by 89
  22. 2025VTool-R1: VLMs Learn to Think with Images via Reinforcement Learning on Multimodal Tool Usecited by 84
  23. 2023LoRAMoE: Alleviate World Knowledge Forgetting in Large Language Models via MoE-Style Plugincited by 82
  24. 2025AgentAuditor: Human-Level Safety and Security Evaluation for LLM Agentscited by 82
  25. 2025Think Silently, Think Fast: Dynamic Latent Compression of LLM Reasoning Chainscited by 81
  26. 2025A Survey on Latent Reasoningcited by 71
  27. 2025AgentRefine: Enhancing Agent Generalization through Refinement Tuningcited by 51
  28. 2025M+: Extending MemoryLLM with Scalable Long-Term Memorycited by 51
  29. 2025Reinforced Latent Reasoning for LLM-based Recommendationcited by 46
  30. 2024SepLLM: Accelerate Large Language Models by Compressing One Segment into One Separatorcited by 41
  31. 2025Seek in the Dark: Reasoning via Test-Time Instance-Level Policy Gradient in Latent Spacecited by 40
  32. 2024Memory3: Language Modeling with Explicit Memorycited by 39
  33. 2025Parametric Retrieval Augmented Generationcited by 38
  34. 2025SoftCoT++: Test-Time Scaling with Soft Chain-of-Thought Reasoningcited by 38
  35. 2024Online Adaptation of Language Models with a Memory of Amortized Contextscited by 35
  36. 2025ATLaS: Agent Tuning via Learning Critical Stepscited by 29
  37. 2025AgentDistill: Training-Free Agent Distillation with Generalizable MCP Boxescited by 28
  38. 2024Deliberation in Latent Space via Differentiable Cache Augmentationcited by 25

50 most cited of 54 All 54 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 4; 2026: 95.

99 papers.

Sort:

By area:

  1. 2026MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agentscited by 129
  2. 2026SkillNet: Create, Evaluate, and Connect AI Skillscited by 48
  3. 2026Skill-Pro: Learning Reusable Skills from Experience via Non-Parametric PPO for LLM Agentscited by 44
  4. 2026A Survey of Agent Memory in the Second Half: Towards Self-Evolving and Long-Horizon Agentscited by 40
  5. 2025VisMem: Latent Vision Memory Unlocks Potential of Vision-Language Modelscited by 31
  6. 2025MemVerse: Multimodal Memory for Lifelong Learning Agentscited by 31
  7. 2025Reasoning Within the Mind: Dynamic Multimodal Interleaving in Latent Spacecited by 25
  8. 2026Graph-based Agent Memory: Taxonomy, Techniques, and Applicationscited by 22
  9. 2026Anatomy of Agentic Memory: Taxonomy and Empirical Analysis of Evaluation and System Limitationscited by 20
  10. 2026From Storage to Experience: A Survey on the Evolution of LLM Agent Memory Mechanismscited by 19
  11. 2026Memory Matters More: Event-Centric Memory as a Logic Map for Agent Searching and Reasoningcited by 17
  12. 2026Self-Improvements in Modern Agentic Systems: A Surveycited by 17
38 more
  1. 2026Learning to Continually Learn via Meta-learning Agentic Memory Designscited by 16
  2. 2026Beyond Dialogue Time: Temporal Semantic Memory for Personalized LLM Agentscited by 14
  3. 2026HyperMem: Hypergraph Memory for Long-Term Conversationscited by 14
  4. 2026MemBuilder: Reinforcing LLMs for Long-Term Memory Construction via Attributed Dense Rewardscited by 13
  5. 2025AI Meets Brain: Memory Systems from Cognitive Neuroscience to Autonomous Agentscited by 12
  6. 2026LatentMem: Customizing Latent Memory for Multi-Agent Systemscited by 12
  7. 2026Hybrid Self-evolving Structured Memory for GUI Agentscited by 12
  8. 2026MedSynapse-V: Bridging Visual Perception and Clinical Intuition via Latent Memory Evolutioncited by 12
  9. 2026δ-mem: Efficient Online Memory for Large Language Modelscited by 12
  10. 2026Toward Efficient Agents: Memory, Tool learning, and Planningcited by 11
  11. 2026Memex(RL): Scaling Long-Horizon LLM Agents via Indexed Experience Memorycited by 11
  12. 2026NextMem: Towards Latent Factual Memory for LLM-based Agentscited by 10
  13. 2026Dual Latent Memory for Visual Multi-agent Systemcited by 9
  14. 2026FlashMem: Distilling Intrinsic Latent Memory via Computation Reusecited by 7
  15. 2026Show, Don't Tell: Morphing Latent Reasoning into Image Generationcited by 7
  16. 2026TAME: A Trustworthy Test-Time Evolution of Agent Memory with Systematic Benchmarkingcited by 7
  17. 2026Position: Modular Memory is the Key to Continual Learning Agentscited by 7
  18. 2026OCR-Memory: Optical Context Retrieval for Long-Horizon Agent Memorycited by 7
  19. 2026Gated Differentiable Working Memory for Long-Context Language Modelingcited by 6
  20. 2026Learning to Remember: End-to-End Training of Memory Agents for Long-Context Reasoningcited by 6
  21. 2026Latent On-Policy Self-Distillationcited by 6
  22. 2026Aligning Agentic World Models via Knowledgeable Experience Learningcited by 5
  23. 2026Large Language Model Agents Are Not Always Faithful Self-Evolverscited by 5
  24. 2026GSEM: Graph-based Self-Evolving Memory for Experience Augmented Clinical Reasoningcited by 5
  25. 2026MSA: Memory Sparse Attention for Efficient End-to-End Memory Model Scaling to 100M Tokenscited by 5
  26. 2026From History to State: Constant-Context Skill Learning for LLM Agentscited by 5
  27. 2026TRUSTMEM: Learning Trustworthy Memory Consolidation for LLM Agents with Long-Term Memorycited by 5
  28. 2026Metis: Memory Foundation Modelcited by 5
  29. 2026SwiftMem: Fast Agentic Memory via Query-aware Indexingcited by 4
  30. 2026HyMem: Hybrid Memory Architecture with Dynamic Retrieval Schedulingcited by 4
  31. 2026MMA: Multimodal Memory Agentcited by 4
  32. 2026MeMo: Memory as a Modelcited by 4
  33. 2026ElasticMem: Latent Memory as a Learnable Resource for LLM Agentscited by 4
  34. 2026Dual Latent Memory in Vision-Language-Action Models for Robotic Manipulationcited by 4
  35. 2026Experience-Driven Multi-Agent Systems Are Training-free Context-aware Earth Observerscited by 3
  36. 2026KEEP: A KV-Cache-Centric Memory Management System for Efficient Embodied Planningcited by 3
  37. 2026Oblivion: Self-Adaptive Agentic Memory Control through Decay-Driven Activationcited by 3
  38. 2026Mem-W: Latent Memory-Native GUI Agentscited by 3

50 most cited of 99 All 99 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026MemWeaver: Weaving Hybrid Memories for Traceable Long-Horizon Agentic Reasoning
  2. 2026MemGym: a Long-Horizon Memory Environment for LLM Agents
  3. 2025Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution
  4. 2026Memory as a Controlled Process: Learned Adaptive Memory Management for LLM Agents
  5. 2025MemEvolve: Meta-Evolution of Agent Memory Systems
  6. 2026SAM: State-Adaptive Memory for Long-Horizon Reasoning Agent
  7. 2026Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents
  8. 2026EvoMemBench: Benchmarking Agent Memory from a Self-Evolving Perspective