Procedural Memory

AdMem unifies Actor agent, Long-term memory agent, Critic agent, Short-term memory, and Reward-based long-term memory management into a bi-level semantic episodic procedural memory system for task-solving agents. On AgentBoard’s AlfWorld domain, AdMem reaches 63.4% task completeness versus 49.3% for ReAct, a +14.1 percentage point gain.

APEX-EM combines a Procedural Knowledge Graph, Experience Memory store, PRGII workflow, Task Verifiers, and StructuralSignatureExtractor to store and reuse full procedural-episodic traces without changing model weights. On KGQAGen-10k, APEX-EM reaches 89.6% accuracy (95.3% CSR) versus 41.3% without memory and surpasses the GPT-4o w/ SP oracle at 84.9%.

ContraMem builds Function Cards, Skill Cards, a Reflector, and a Curator that turn heterogeneous same-task trajectories into compact procedural memory. On GAIA2/ARE, ContraMem raises the source-target macro success rate from 26.2% with no memory to 55.3%, 30.0 points above Agent Workflow Memory.

HYPERSKILL organizes experiential memory as a hypergraph of Nodes, Hyperedges, Dual-path retrieval, Co-occurrence ranking, and HYPERSKILL Memory Evolving to store and reuse subtask and skill knowledge. On GAIA with GPT-4o, HYPERSKILL achieves 44.24% success rate, improving over PlugMem’s 41.94% by +2.30 points.

MindMemOS organizes memories in a unified entity–property–time graph, with MindVanilla, MindSchema, MindMemEvolve, dreaming, feedback, and MindSkillEvolve coordinating memory and skill evolution. MindMemOS reaches 94.03% overall accuracy on LOCOMO and 70.63% on PersonaMem, surpassing EverOS by 0.98 and 3.06 percentage points respectively.

SkillEvolBench orchestrates Environment-specific Shared Skill Library, Skill Author, Episodic Task Attempts, and Structured Verifier Feedback to turn trajectories into external procedural skills and then freeze them. SkillEvolBench’s main result is that Raw-Trajectory achieves mean ESR 37.6% and ARSR 44.7%, while most curated and self-generated skill variants fail to match these deployment scores.

CodeMem reframes agents as workflow architects that combine search_functions, write_todos, execute_code, and register_skill to build deterministic, reusable Python skills. On a 25-task benchmark, CodeMem with Gemini 3 Full achieves 96% minimum correctness with 7.00 calls and 2.02M tokens, compared to 68% correctness and 2.80 calls for GPT-5 Chat.

Evaluating Long-Term Memory for Long-Context Question Answering compares Full Context, RAG, A-Mem, RAG+PromptOpt, and RAG+EpMem memory components across semantic, episodic, and procedural memory for long conversational QA. On LoCoMo, RAG+EpMem reaches an average F1 ranking of 1.83 for Llama 3.2-3B Instruct and 1.80 for GPT-4o mini while using around 1,000 tokens per query versus over 23,000 for Full Context.

LEGOMem builds a procedural memory bank, full task memories, subtask memories, LEGOMem-Dynamic, and LEGOMem-QueryRewrite to store successful trajectories and allocate them to orchestrators and task agents. On OfficeBench, LEGOMem reaches 58.44% overall success for the LLM team versus 45.83% with no memory, and lifts the SLM team from 24.78% to 38.16%.

Memp constructs agent skills via Build, Retrieve, and Update modules that turn past trajectories into scripts, trajectories, and combined proceduralizations stored in a procedural memory library. On ALFWorld, Memp’s proceduralization with GPT-4o reaches 77.86% test success versus 42.14% with no memory, while reducing steps from 23.76 to 15.01.

ReMe manages procedural memory through experience acquisition, experience reuse, and experience refinement, combining multi-faceted distillation, context-adaptive reuse, and utility-based deletion into a single lifecycle. On BFCL-V3 and AppWorld, Qwen3-8B with ReMe (dynamic) achieves 34.94% Avg@4 vs 27.65% for the No Memory baseline, and 55.03% Pass@4 vs 46.20%, showing that self-evolving memory can substitute for model scale.

TokMem adds a Memory Bank of trainable memory tokens to a frozen Transformer backbone, using memory routing, conditional generation, and renormalization to store and recall procedures. On Super-Natural Instructions and APIGen function-calling, TokMem reaches 67.0 ROUGE-L and 99.1 tool-selection F1, surpassing Replay Memory and LoRA fine-tuning with far fewer parameters.