memory_mechanism

Memory systems and mechanisms in AI

Name	Categories	Topics	References	NavigationOrder	Eleventy	TemplateEngineOverride	Credibility	Title	RelatedNotes
20260411_dreaming	self_dream	memory_mechanism	claw	[object Object]	[object Object]	md
20260414_gems_agent_native_multimodal_generation_memory_skills	paper_analysis	agent_architecture, multimodal, memory_mechanism, long_term_memory, llm	shanghai_ai_lab, sjtu	[object Object]	[object Object]	md	100	GEMS: Agent-Native Multimodal Generation with Memory and Skills
20260419_memskill_learning_evolving_memory_skills	paper_analysis	self_evolving_agents, memory_mechanism, agent_architecture, llm, long_term_memory	nanyang_technological_university, tsinghua_university, university_of_illinois_urbana_champaign	[object Object]	[object Object]	md	100	MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agents	GEMS: Agent-Native Multimodal Generation with Memory and Skills
20260422_dreaming	self_dream, self_insight	memory_mechanism, symbolic_reasoning	claw	[object Object]	[object Object]	md
20260422_hipporag_neurobiologically_inspired_long_term_memory	paper_analysis	rag, knowledge_graph, multi_hop_reasoning, neuro_science, memory_mechanism	personalized_pagerank, musique, hotpotqa, 2wiki, natural_questions	[object Object]	[object Object]	md	100	HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models
20260422_memory_os_of_ai_agent	paper_analysis	agent_architecture, memory_mechanism, long_term_memory, context_engineering	locomo_bench, bai_lab	[object Object]	[object Object]	md	100	Memory OS of AI Agent	HippoRAG: Neurobiologically Inspired Long-Term Memory, M3-Agent: Multimodal Agent with Long-Term Memory
20260422_multimodal_agent_long_term_memory	paper_analysis	agent_architecture, multimodal, memory_mechanism, long_term_memory, video_understanding	m3_bench, bytedance_seed, dapo	[object Object]	[object Object]	md	100	Seeing, Listening, Remembering, and Reasoning: A Multimodal Agent with Long-Term Memory	HippoRAG: Neurobiologically Inspired Long-Term Memory
20260425_reca_integrated_acceleration_cooperative_embodied_agents	paper_analysis	llm, agent_architecture, memory_mechanism, embodied_ai, multi_hop_reasoning	georgia_tech, harvard_university	[object Object]	[object Object]	md	100	ReCA: Integrated Acceleration for Real-Time and Efficient Cooperative Embodied Autonomous Agents
20260427_agentsquare_automatic_llm_agent_search	paper_analysis	agent_architecture, llm, reasoning, memory_mechanism, reinforce_learning	tsinghua_fib_lab, neural_architecture_search	[object Object]	[object Object]	md	100	AgentSquare: Automatic LLM Agent Search in Modular Design Space
20260427_detecting_hallucinations_semantic_entropy	paper_analysis	llm, reasoning, agent_architecture, memory_mechanism, reinforce_learning	oatml_oxford, neural_architecture_search	[object Object]	[object Object]	md	100	Detecting hallucinations in large language models using semantic entropy
20260429_agent_workflow_memory	paper_analysis	agent_architecture, memory_mechanism, llm, gui_agent, long_term_memory, cognitive_science	cmu, mit, mind2web, webarena	[object Object]	[object Object]	md	100	Agent Workflow Memory	disentangling_memory_reasoning_llm
20260429_disentangling_memory_reasoning_llm	paper_analysis	memory_mechanism, reasoning, llm, multi_hop_reasoning, chain_of_thought, cognitive_science	rutgers_university, ohio_state_university, ucsb, truthfulqa, strategyqa, commonsenseqa	[object Object]	[object Object]	md	100	Disentangling Memory and Reasoning Ability in Large Language Models	agent_workflow_memory
20260430_catastrophic_forgetting_implicit_inference	paper_analysis	catastrophic_forgetting, llm, memory_mechanism, reasoning, context_engineering	cmu, conjugate_prompting	[object Object]	[object Object]	md	100	Understanding Catastrophic Forgetting in Language Models via Implicit Inference
20260504_reasoningbank_scaling_agent_self_evolving_reasoning_memory	paper_analysis	self_evolving_agents, reasoning_memory, test_time_scaling, agent_architecture, memory_mechanism	google_cloud_ai_research, webarena, mind2web, swebench_verified	[object Object]	[object Object]	md	100	ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory	Self-Evolving Agents: A Survey
20260504_self_evolving_agents_survey_asi	paper_analysis	self_evolving_agents, artificial_super_intelligence, survey, agent_architecture, memory_mechanism	princeton_university, tsinghua_university, sjtu, google_cloud_ai_research	[object Object]	[object Object]	md	100	A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence	ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory
20260508_memgpt_towards_llms_as_operating_systems	paper_analysis	llm, memory_mechanism, agent_architecture, long_term_memory, context_engineering	uc_berkeley, memgpt	[object Object]	[object Object]	md	100	MemGPT: Towards LLMs as Operating Systems
20260509_skillos_learning_skill_curation_self_evolving_agents	paper_analysis	agent_architecture, memory_mechanism, reinforce_learning, self_evolving_agents, skill_curation	google_cloud_ai_research, grpo, skillos, alfworld	[object Object]	[object Object]	md	100	SkillOS: Learning Skill Curation for Self-Evolving Agents	self_evolving_agents_survey_asi, memory_os_of_ai_agent, agentevolver_self_evolving_agent, reasoningbank_scaling_agent_self_evolving_reasoning_memory
20260512_gasim_graph_accelerated_hybrid_social_simulation	paper_analysis	agent_architecture, memory_mechanism, knowledge_graph, llm, multi_agent_systems	uestc, graph_attention_network, mem0	[object Object]	[object Object]	md	100	GASim: A Graph-Accelerated Hybrid Framework for Social Simulation	memgpt_towards_llms_as_operating_systems
20260512_learning_fast_slow_llms_adapt_continually	paper_analysis	catastrophic_forgetting, continual_learning, memory_mechanism, reinforce_learning, llm	uc_berkeley	[object Object]	[object Object]	md	100	Learning, Fast and Slow: Towards LLMs That Adapt Continually	20260504_reasoningbank_scaling_agent_self_evolving_reasoning_memory, 20260511_rl_long_horizon_reasoning_llm_expressiveness
20260512_meme_multi_entity_evolving_memory_evaluation	paper_analysis	memory_mechanism, long_term_memory, agent_architecture, llm, reasoning, evaluation	kaist_ai, tuebingen_ai_center, naver_ai_lab	[object Object]	[object Object]	md	100	MEME: Multi-entity & Evolving Memory Evaluation	memory_os_of_ai_agent, memgpt_towards_llms_as_operating_systems, disentangling_memory_reasoning_llm
20260515_harnessing_agentic_evolution	paper_analysis	agent_architecture, self_evolving_agents, reasoning, memory_mechanism, reinforce_learning, workflow_optimization	deepwisdom, hkust_gz, sjtu, tsinghua_university, nanyang_technological_university	[object Object]	[object Object]	md	100	harnessing_agentic_evolution	self_evolving_agents_survey_asi, skillos_learning_skill_curation_self_evolving_agents, aflow_automating_agentic_workflow_generation
20260517_heterogeneous_temporal_memory_governance_llm_persona	paper_analysis	memory_mechanism, llm, long_term_memory, rag, reasoning, persona_consistency	uestc, arpm	[object Object]	[object Object]	md	100	A Heterogeneous Temporal Memory Governance Framework for Long-Term LLM Persona Consistency	memgpt_towards_llms_as_operating_systems, meme_multi_entity_evolving_memory_evaluation
20260518_memlineage_lineage_guided_llm_agent_memory	paper_analysis	memory_mechanism, agent_architecture, llm, agent_security, lineage_tracking	iie_cas	[object Object]	[object Object]	md	100	MemLineage: Lineage-Guided Enforcement for LLM Agent Memory
20260522_calmem_dual_memory_conversational_ai	paper_analysis	memory_mechanism, long_term_memory, agent_architecture, rag, context_engineering	infosys_limited, memgpt	[object Object]	[object Object]	md	100	Application-Layer Dual Memory for Conversational AI: Achieving Virtually Unbounded Context Without Model Modification	memgpt_towards_llms_as_operating_systems, memlineage_lineage_guided_llm_agent_memory
20260525_gated_deltanet_2_decoupling_erase_write_linear_attention	paper_analysis	memory_mechanism, llm, reasoning, long_term_memory, reinforce_learning	nvidia, deltanet	[object Object]	[object Object]	md	100	Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention
20260527_scaling_harness_agentic_ai	paper_analysis	agent_architecture, memory_mechanism, reasoning, self_evolving_agents, multi_agent_systems	uc_berkeley, cheetahclaws	[object Object]	[object Object]	md	100	From Model Scaling to System Scaling: Scaling the Harness in Agentic AI	MemLineage, MemGPT, SkillOS, CalMem
20260528_CORE_contrastive_reflection_reasoning	paper_analysis	reasoning, memory_mechanism, self_evolving_agents, contrastive_reflection, reinforce_learning, cognitive_science	stanford_iris_lab, grpo, memgpt	[object Object]	[object Object]	md	100	CORE: Contrastive Reflection Enables Rapid Improvements in Reasoning	Can RL Teach Long-Horizon Reasoning to LLMs? Expressiveness Is Key, MemLineage: Lineage-Guided Enforcement for LLM Agent Memory, AMR-SD: Asymmetric Meta-Reflective Self-Distillation for Token-Level Credit Assignment
20260528_muse_autoskill_self_evolving_skill_memory	paper_analysis	agent_architecture, memory_mechanism, self_evolving_agents, skill_curation, continual_learning	bytedance_seed, skillsbench, muse_autoskill	[object Object]	[object Object]	md	100	MUSE-Autoskill: Self-Evolving Agents via Skill Creation, Memory, Management, and Evaluation	skillos_learning_skill_curation_self_evolving_agents, skillopt_executive_strategy_self_evolving_agent_skills, agent_memory_and_reasoning_frontier_survey_2026
20260528_unlocking_working_memory_latent_reasoning	paper_analysis	memory_mechanism, reasoning, llm, reasoning_memory, context_engineering	lit_ai_lab, nxai_gmbh, gsm8k	[object Object]	[object Object]	md	100	Unlocking the Working Memory of Large Language Models for Latent Reasoning	chain_of_thought_prompting_elicits_reasoning, memgpt_towards_llms_as_operating_systems, agent_memory_and_reasoning_frontier_survey_2026
20260529_agent_lifespan_engineering_for_deployed_systems	paper_analysis	memory_mechanism, long_term_memory, agent_architecture, evaluation, self_evolving_agents	ut_austin	[object Object]	[object Object]	md	100	Your Agents Are Aging Too: Agent Lifespan Engineering for Deployed Systems	Your Agents Are Aging Too: Agent Lifespan Engineering for Deployed Systems
20260602_lintree_explicitly_structured_search_histories	paper_analysis\|paper_analysis	reasoning, llm, agent_architecture, memory_mechanism	nus, oatml_oxford	[object Object]	[object Object]	md	100	LinTree: Improving LLM Reasoning with Explicitly Structured Search Histories	chain_of_thought_prompting_elicits_reasoning
20260603_agentcl_continual_learning_language_agents	paper_analysis	continual_learning, memory_mechanism, agent_architecture, evaluation, self_evolving_agents	agentcl, ohio_state_university	[object Object]	[object Object]	md	100	AGENTCL: Toward Rigorous Evaluation of Continual Learning in Language Agents	learning_fast_slow_llms_adapt_continually, meme_multi_entity_evolving_memory_evaluation, muse_autoskill_self_evolving_skill_memory
20260604_agent_memory_characterization_system_implications	paper_analysis	memory_mechanism, agent_architecture, long_term_memory, llm, evaluation	stanford_university, ku_leuven, memory_agent_bench, memgpt, graphrag	[object Object]	[object Object]	md	95	Agent Memory: Characterization and System Implications of Stateful Long-Horizon Workloads	memgpt_towards_llms_as_operating_systems, memlineage_lineage_guided_llm_agent_memory, calmem_dual_memory_conversational_ai, unlocking_working_memory_latent_reasoning, aip_graph_representation_for_learning_and_governing_agent_skills
20260604_language_models_need_sleep_self_modify_consolidate_memories	paper_analysis	memory_mechanism, self_evolving_agents, continual_learning, reinforce_learning, reasoning, llm	google_brain, knowledge_seeding, sleep_paradigm	[object Object]	[object Object]	md	100	Language Models Need Sleep: Learning to Self-Modify and Consolidate Memories	memory_os_of_ai_agent, hipporag_neurobiologically_inspired_long_term_memory, muse_autoskill_self_evolving_skill_memory
20260604_pretraining_recurrent_networks_without_recurrence	paper_analysis	memory_mechanism, recurrent_neural_networks, llm, reinforce_learning, neuro_science	mit, supervised_memory_training, backpropagation_through_time	[object Object]	[object Object]	md	100	Pretraining Recurrent Networks without Recurrence
20260609_socratic_swe_self_evolving_coding_agents_trace_derived_skills	paper_analysis	self_evolving_agents, memory_mechanism, skill_curation, reinforce_learning, reasoning	sjtu, swebench_verified, terminalbench_2, grpo, skillos	[object Object]	[object Object]	md	100	Socratic-SWE: Self-Evolving Coding Agents via Trace-Derived Agent Skills	skillos_learning_skill_curation_self_evolving_agents, skillopt_executive_strategy_self_evolving_agent_skills, aflow_automating_agentic_workflow_generation, scaling_harness_agentic_ai, meme_multi_entity_evolving_memory_evaluation
20260617_evolvenav_proactive_preflection_self_evolving_memory	paper_analysis	embodied_ai, semantic_navigation, self_evolving_agents, llm, memory_mechanism	hkust_gz, nanyang_technological_university, habitat_simulator	[object Object]	[object Object]	md	100	EvolveNav: Proactive Preflection and Self-Evolving Memory for Zero-Shot Object Goal Navigation
20260618_fixed_point_reasoners_stable_adaptive_deep_looped_transformers	paper_analysis	reasoning, llm, recurrent_neural_networks, memory_mechanism, agent_architecture	tuebingen_ai_center, eth_zurich, liquidai, max_planck_institute	[object Object]	[object Object]	md	100	Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformers
20260624_evoarena_tracking_memory_evolution	paper_analysis	memory_mechanism, agent_architecture, llm, evaluation, embodied_ai	nus, ntu, terminalbench_2	[object Object]	[object Object]	md	100	EvoArena: Tracking Memory Evolution for Robust LLM Agents in Dynamic Environments	Self-Harness: Harnesses That Improve Themselves
20260627_joint_learning_experiential_rules_policies_llm_agents	paper_analysis	memory_mechanism, reinforce_learning, self_evolving_agents, reasoning, agent_architecture	alfworld, grpo, sun_yat_sen_university	[object Object]	[object Object]	md	100	Joint Learning of Experiential Rules and Policies for Large Language Model Agents	unlocking_working_memory_latent_reasoning, memory_os_of_ai_agent, memskill_learning_evolving_memory_skills
20260629_carve_content_aware_recurrent_value_efficiency	paper_analysis	memory_mechanism, recurrent_neural_networks, llm, reasoning, linear_attention	carve, deltanet	[object Object]	[object Object]	md	100	CARVE: Content-Aware Recurrent with Value Efficiency for Chunk-Parallel Linear Attention	Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention
20260629_memory_r1_enhancing_llm_agents_manage_utilize_memories_rl	paper_analysis	agent_architecture, memory_mechanism, reinforce_learning, llm, long_term_memory	lmu_munich, mem0, locomo_bench, grpo	[object Object]	[object Object]	md	100	Memory-R1: Enhancing Large Language Model Agents to Manage and Utilize Memories via Reinforcement Learning	memgpt_towards_llms_as_operating_systems, muse_autoskill_self_evolving_skill_memory, gems_agent_native_multimodal_generation_memory_skills, amr_sd_token_level_credit_assignment, agent_memory_characterization_system_implications
20260630_from_tokens_to_states_llms_world_models	paper_analysis	llm, reasoning, world_model, memory_mechanism, agent_architecture	jepa, othello_gpt	[object Object]	[object Object]	md	100	From Tokens to States: LLMs as a Special Case of World Models and the Continuous Path Beyond	memory_os_of_ai_agent, rl_long_horizon_reasoning_llm_expressiveness
20260702_memory_in_the_age_of_ai_agents_a_survey	paper_analysis	memory_mechanism, agent_architecture, survey, llm, long_term_memory	nus, fudan_university, pku, renmin_university_of_china	[object Object]	[object Object]	md	100	Memory in the Age of AI Agents: A Survey — Forms, Functions and Dynamics
20260703_self_evolving_world_models_llm_agent_planning	paper_analysis	world_model, self_evolving_agents, memory_mechanism, embodied_ai, reasoning	nus, alfworld, deepmind	[object Object]	[object Object]	md	100	Self-Evolving World Models for LLM Agent Planning	from_tokens_to_states_llms_world_models, world_models_in_pieces_structural_certification, memory_r1_enhancing_llm_agents_manage_utilize_memories_rl, stop_comparing_llm_agents_without_disclosing_harness
20260710_loger_long_context_geometric_reconstruction	paper_analysis	long_term_memory, memory_mechanism, embodied_ai, multimodal, deep_learning	loger, google_deepmind, kitti, vbr	[object Object]	[object Object]	md	100	LoGeR: Long-Context Geometric Reconstruction with Hybrid Memory
20260713_metacognition_in_llms_foundations_progress_opportunities	paper_analysis, domain_survey	llm, reasoning, cognitive_science, memory_mechanism, agent_architecture	yale_university, uc_irvine	[object Object]	[object Object]	md	100	Metacognition in LLMs: Foundations, Progress, and Opportunities	chain_of_thought_prompting_elicits_reasoning, ai_reasoning_deep_learning_symbolic_neural, memory_os_of_ai_agent
20260713_shared_selective_persistent_memory_agentic_llm	paper_analysis	agent_architecture, memory_mechanism, long_term_memory, multi_agent_systems, context_engineering	apple_inc, graygoo	[object Object]	[object Object]	md	100	Shared Selective Persistent Memory for Agentic LLM Systems	memgpt_towards_llms_as_operating_systems, agent_memory_characterization_system_implications, memory_in_the_age_of_ai_agents_a_survey
20260717_experience_memory_graph_one_shot_error_correction_for_agents	paper_analysis	memory_mechanism, agent_architecture, reasoning, long_term_memory, knowledge_graph	alfworld, scienceworld, uestc	[object Object]	[object Object]	md	100	Experience Memory Graph: One-Shot Error Correction for Agents	memlineage_lineage_guided_llm_agent_memory, shared_selective_persistent_memory_agentic_llm, agent_memory_and_reasoning_frontier_survey_2026
20260719_t2mlr_transformer_temporal_middle_layer_recurrence	paper_analysis	reasoning, llm, memory_mechanism, recurrent_neural_networks, reasoning_memory	princeton_university, chain_of_thought, backpropagation_through_time, gsm8k, hotpotqa	[object Object]	[object Object]	md	100	T²MLR: Transformer with Temporal Middle-Layer Recurrence	searchos_v1_open_domain_information_seeking_agent_collaboration, chain_of_thought_prompting_elicits_reasoning, memory_r1_enhancing_llm_agents_manage_utilize_memories_rl
20260722_supra_cognitive_modes_routed_agent_memory	paper_analysis	agent_architecture, memory_mechanism, long_term_memory, reasoning, multi_hop_reasoning	supra_research, locomo_bench, memory_agent_bench, long_mem_eval	[object Object]	[object Object]	md	100	Supra Cognitive Modes: A Routed Architecture for Agent Memory	memory_os_of_ai_agent, memgpt_towards_llms_as_operating_systems, shared_selective_persistent_memory_agentic_llm
20260724_pro_long_programmatic_memory_long_horizon_reasoning	paper_analysis	memory_mechanism, agent_architecture, reasoning, long_term_memory, coding_agents, continual_learning	duke_university, arc_agi, chain_of_thought	[object Object]	[object Object]	md	100	PRO-LONG: Programmatic Memory Enables Long-Horizon Reasoning
20260726_agentic_context_management_memory_cost_lifecycle_architecture	paper_analysis	agent_architecture, memory_mechanism, context_engineering, reasoning, long_term_memory	maximem, long_mem_eval, locomo_bench	[object Object]	[object Object]	md	100	Agentic Context Management: Solving Agent Memory and Cost by Treating Them as Lifecycle and Architecture Problems	memory_os_of_ai_agent, agent_memory_and_reasoning_frontier_survey_2026, memory_in_the_age_of_ai_agents_a_survey, memgpt_towards_llms_as_operating_systems, context_engineering_2_overview
20260731_memrl_self_evolving_agents_runtime_rl_episodic_memory	paper_analysis	memory_mechanism, reinforce_learning, agent_architecture, llm, reasoning, long_term_memory	sjtu, xidian_university, nus, shanghai_innovation_institute, memtensor	[object Object]	[object Object]	md	100	MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic Memory