# Learning Hierarchical Procedural Memory for LLM Agents through Bayesian Selection and Contrastive Refinement

arXiv: https://arxiv.org/abs/2512.18950 · HTML: https://mlsanity.com/paper/2512.18950/

## Key facts

- Cited by 21 papers in the map (data through 14 Oct 2026).
- Placed in Agent memory systems, one of 16 directions in LLM agents and tool use.
- Builds on 16 of its 28 references that are in the map.
- First posted on arXiv 22 Dec 2025.

## Placed in

- LLM agents and tool use / [Agent memory systems](https://mlsanity.com/agents/agent-memory-systems/)

## Builds on (most cited first, 16 of 16)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [ALFWorld: Aligning Text and Embodied Environments for Interactive Learning](https://mlsanity.com/paper/2010.03768/) (2020)
- [A Survey on the Memory Mechanism of Large Language Model based Agents](https://mlsanity.com/paper/2404.13501/) (2024)
- [TravelPlanner: A Benchmark for Real-World Planning with Language Agents](https://mlsanity.com/paper/2402.01622/) (2024)
- [AgentTuning: Enabling Generalized Agent Abilities for LLMs](https://mlsanity.com/paper/2310.12823/) (2023)
- [FireAct: Toward Language Agent Fine-tuning](https://mlsanity.com/paper/2310.05915/) (2023)
- [MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agent](https://mlsanity.com/paper/2507.02259/) (2025)
- [InterCode: Standardizing and Benchmarking Interactive Coding with Execution Feedback](https://mlsanity.com/paper/2306.14898/) (2023)
- [Trial and Error: Exploration-Based Trajectory Optimization for LLM Agents](https://mlsanity.com/paper/2403.02502/) (2024)
- [Watch Every Step! LLM Agent Learning via Iterative Step-Level Process Refinement](https://mlsanity.com/paper/2406.11176/) (2024)
- [AutoManual: Constructing Instruction Manuals by LLM Agents via Interactive Environmental Learning](https://mlsanity.com/paper/2405.16247/) (2024)

## Built on by (most cited first, 21 of 21)

- [MARS: Modular Agent with Reflective Search for Automated AI Research](https://mlsanity.com/paper/2602.02660/) (2026)
- [From Storage to Experience: A Survey on the Evolution of LLM Agent Memory Mechanisms](https://mlsanity.com/paper/2605.06716/) (2026)
- [The Evolution of Tool Use in LLM Agents: From Single-Tool Call to Multi-Tool Orchestration](https://mlsanity.com/paper/2603.22862/) (2026)
- [Generate, Filter, Control, Replay: A Comprehensive Survey of Rollout Strategies for LLM Reinforcement Learning](https://mlsanity.com/paper/2605.02913/) (2026)
- [SuperLocalMemory: Privacy-Preserving Multi-Agent Memory with Bayesian Trust Defense Against Memory Poisoning](https://mlsanity.com/paper/2603.02240/) (2026)
- [Ratchet: How Reliable Must an LLM Judge Be to Retire a Skill?](https://mlsanity.com/paper/2605.22148/) (2026)
- [Security Considerations for Multi-agent Systems](https://mlsanity.com/paper/2603.09002/) (2026)
- [Oblivion: Self-Adaptive Agentic Memory Control through Decay-Driven Activation](https://mlsanity.com/paper/2604.00131/) (2026)
- [AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents](https://mlsanity.com/paper/2607.02255/) (2026)
- [View-oriented Conversation Compiler for Agent Trace Analysis](https://mlsanity.com/paper/2603.29678/) (2026)
- [AdaptFuse: Training-Free Sequential Preference Inference via Externalized Bayesian Computation](https://mlsanity.com/paper/2604.03925/) (2026)
- [Web2BigTable: A Bi-Level Multi-Agent LLM System for Internet-Scale Information Search and Extraction](https://mlsanity.com/paper/2604.27221/) (2026)
- [Skill-CMIB: Multimodal Agent Skill for Consistent Action via Conditional Multimodal Information Bottleneck](https://mlsanity.com/paper/2605.08526/) (2026)
- [AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Composition](https://mlsanity.com/paper/2606.14674/) (2026)
- [Making AI Scientists Auditable from Evidence to Claim](https://mlsanity.com/paper/2606.18874/) (2026)
- [Designer-RSI: Evolving Procedural Memory from User Traffic for Agentic Graphic Design](https://mlsanity.com/paper/2609.22086/) (2026)
- [CLEAR: Context Augmentation from Contrastive Learning of Experience via Agentic Reflection](https://mlsanity.com/paper/2604.07487/) (2026)
- [Task-Adaptive Retrieval over Agentic Multi-Modal Web Histories via Learned Graph Memory](https://mlsanity.com/paper/2604.07863/) (2026)
- [Personalization Meets Safety:Mechanisms,Risks,and Mitigations in Personalized LLMs](https://mlsanity.com/paper/2606.09038/) (2026)
- [Activity Frames: Deterministic Screen-Activity Compilation for Agent Memory and Replay](https://mlsanity.com/paper/2608.05784/) (2026)
- [AECG: Asymmetric Experience Consolidation and Governance In Multi-Agent Systems](https://mlsanity.com/paper/2610.05176/) (2026)

## Abstract (arXiv)

We present MACLA, a framework that decouples reasoning from learning by maintaining a frozen large language model while performing all adaptation in an external hierarchical procedural memory. MACLA extracts reusable procedures from trajectories, tracks reliability via Bayesian posteriors, selects actions through expected-utility scoring, and refines procedures by contrasting successes and failures. Across four benchmarks (ALFWorld, WebShop, TravelPlanner, InterCodeSQL), MACLA achieves 78.1 percent average performance, outperforming all baselines. On ALFWorld unseen tasks, MACLA reaches 90.3 percent with 3.1 percent positive generalization. The system constructs memory in 56 seconds, 2800 times faster than the state-of-the-art LLM parameter-training baseline, compressing 2851 trajectories into 187 procedures. Experimental results demonstrate that structured external memory with Bayesian selection and contrastive refinement enables sample-efficient, interpretable, and continually improving agents without LLM parameter updates.