# Agentic Recommender System with Hierarchical Belief-State Memory

arXiv: https://arxiv.org/abs/2605.14401 · HTML: https://mlsanity.com/paper/2605.14401/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Agentic recommender systems, one of 16 directions in LLM agents and tool use.
- Builds on 36 of its 43 references that are in the map.
- First posted on arXiv 14 May 2026.

## Placed in

- LLM agents and tool use / [Agentic recommender systems](https://mlsanity.com/agents/agentic-recommender-systems/)

## Builds on (most cited first, 30 of 36)

- [MemGPT: Towards LLMs as Operating Systems](https://mlsanity.com/paper/2310.08560/) (2023)
- [Self-Attentive Sequential Recommendation](https://mlsanity.com/paper/1808.09781/) (2018)
- [LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation](https://mlsanity.com/paper/2002.02126/) (2020)
- [Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory](https://mlsanity.com/paper/2504.19413/) (2025)
- [MemoryBank: Enhancing Large Language Models with Long-Term Memory](https://mlsanity.com/paper/2305.10250/) (2023)
- [Cognitive Architectures for Language Agents](https://mlsanity.com/paper/2309.02427/) (2023)
- [Zep: A Temporal Knowledge Graph Architecture for Agent Memory](https://mlsanity.com/paper/2501.13956/) (2025)
- [Deep Learning Recommendation Model for Personalization and Recommendation Systems](https://mlsanity.com/paper/1906.00091/) (2019)
- [Large Language Models are Zero-Shot Rankers for Recommender Systems](https://mlsanity.com/paper/2305.08845/) (2023)
- [Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations](https://mlsanity.com/paper/2402.17152/) (2024)
- [Recommendation as Instruction Following: A Large Language Model Empowered Recommendation Approach](https://mlsanity.com/paper/2305.07001/) (2023)
- [Memory OS of AI Agent](https://mlsanity.com/paper/2506.06326/) (2025)
- [User Behavior Simulation with Large Language Model based Agents](https://mlsanity.com/paper/2306.02552/) (2023)
- [RecBole: Towards a Unified, Comprehensive and Efficient Framework for Recommendation Algorithms](https://mlsanity.com/paper/2011.01731/) (2020)
- [MemOS: A Memory OS for AI System](https://mlsanity.com/paper/2507.03724/) (2025)
- [Recommender AI Agent: Integrating Large Language Models for Interactive Recommendations](https://mlsanity.com/paper/2308.16505/) (2023)
- [RecMind: Large Language Model Powered Agent For Recommendation](https://mlsanity.com/paper/2308.14296/) (2023)
- [AgentCF: Collaborative Learning with Autonomous Language Agents for Recommender Systems](https://mlsanity.com/paper/2310.09233/) (2023)
- [On Generative Agents in Recommendation](https://mlsanity.com/paper/2310.10108/) (2023)
- [Fine-Grained Spoiler Detection from Large-Scale Review Corpora](https://mlsanity.com/paper/1905.13416/) (2019)
- [LLM-Rec: Personalized Recommendation via Prompting Large Language Models](https://mlsanity.com/paper/2307.15780/) (2023)
- [MACRec: a Multi-Agent Collaboration Framework for Recommendation](https://mlsanity.com/paper/2402.15235/) (2024)
- [Let Me Do It For You: Towards LLM Empowered Recommendation via Tool Learning](https://mlsanity.com/paper/2405.15114/) (2024)
- [A Survey on LLM-powered Agents for Recommender Systems](https://mlsanity.com/paper/2502.10050/) (2025)
- [PersonaX: A Recommendation Agent Oriented User Modeling Framework for Long Behavior Sequence](https://mlsanity.com/paper/2503.02398/) (2025)
- [SimUSER: Simulating User Behavior with Large Language Models for Recommender System Evaluation](https://mlsanity.com/paper/2504.12722/) (2025)
- [iAgent: LLM Agent as a Shield between User and Recommender Systems](https://mlsanity.com/paper/2502.14662/) (2025)
- [MemRec: Collaborative Memory-Augmented Agentic Recommender System](https://mlsanity.com/paper/2601.08816/) (2026)
- [RecUserSim: A Realistic and Diverse User Simulator for Evaluating Conversational Recommender Systems](https://mlsanity.com/paper/2507.22897/) (2025)
- [Partially Observable Markov Decision Process for Recommender Systems](https://mlsanity.com/paper/1608.07793/) (2016)

## Built on by (most cited first, 5 of 5)

- [CORAL: An LLM-Native Harness for Production Recommender Systems](https://mlsanity.com/paper/2609.02730/) (2026)
- [Do Recommendation Algorithms Work When Users Are LLM Agents? A Case Study on Moltbook](https://mlsanity.com/paper/2606.29762/) (2026)
- [Dear Algo: A Precision-First Agentic Intent Layer for Unified Search and Recommendation](https://mlsanity.com/paper/2608.15877/) (2026)
- [PersonaMem-v3: Toward Omni-Platform Personal Intelligence for Holistic User Understanding, Recommendation, and Agentic Tasks](https://mlsanity.com/paper/2608.21381/) (2026)
- [Personal-Agent Mediated Recommendation with Cross-Platform User History](https://mlsanity.com/paper/2610.07588/) (2026)

## Abstract (arXiv)

Memory-augmented LLM agents have advanced personalized recommendation, yet existing approaches universally adopt flat memory representations that conflate ephemeral signals with stable preferences, and none provides a complete lifecycle governing how memory should evolve. We propose MARS (Memory-Augmented Agentic Recommender System), a framework that treats recommendation as a partially observable problem and maintains a structured belief state that progressively abstracts noisy behavioral observations into a compact estimate of user preferences. MARS organizes this belief state into three tiers: event memory buffers raw signals, preference memory maintains fine-grained mutable chunks with explicit strength and evidence tracking, and profile memory distills all preferences into a coherent natural language narrative. A complete lifecycle of six operations -- extraction, reinforcement, weakening, consolidation, forgetting, and resynthesis -- is adaptively scheduled by an LLM-based planner rather than fixed-interval heuristics. Experiments on four InstructRec benchmark domains show that MARS achieves state-of-the-art performance with average improvements of 26.4% in HR@1 and 10.3% in NDCG@10 over the strongest baselines with further gains from agentic scheduling in evolving settings.