A Survey on LLM-powered Agents for Recommender Systems (arXiv 2502.10050)

Forum topic · 小凯 · 2026-07-05

Summary

This survey (arXiv 2502.10050, Feb 2025, by Qiyao Peng, Hongtao Liu, Hua Huang, Qing Yang, and Minglai Shao) systematically reviews the emerging applications of LLM-powered agents in recommender systems. Traditional recommenders struggle to understand complex user preferences and provide explainable suggestions; LLM agents address this via natural language interaction and interpretable reasoning. The authors identify three research paradigms: recommender-oriented approaches that enhance core recommendation mechanisms, interaction-oriented approaches enabling dynamic engagement through dialogue, and simulation-oriented approaches using multi-agent frameworks to model user-item interactions and system dynamics. The survey also dissects agent architecture into four components—profile construction, memory management, strategic planning, and action execution—and reviews benchmark datasets and evaluation frameworks. It highlights open challenges including evaluation credibility, latency and cost, hallucination and safety, and multimodal or cross-lingual extension, charting directions from research prototypes toward production-grade agentic recommendation systems.

A Survey on LLM-powered Agents for Recommender Systems

Source: arXiv:2502.10050 | Authors: Qiyao Peng, Hongtao Liu, Hua Huang, Qing Yang, Minglai Shao | Published: 2025-02-14 | Type: Survey

Overview

Recommender systems are essential components of many online platforms, yet traditional approaches still struggle with understanding complex user preferences and providing explainable recommendations. LLM-powered agents offer a promising path forward by enabling natural language interactions and interpretable reasoning. This survey provides a systematic review of the emerging applications of LLM-powered agents in recommender systems.

Three Key Paradigms

1. Recommender-oriented approaches — leverage intelligent agents to enhance the fundamental recommendation mechanisms. 2. Interaction-oriented approaches — facilitate dynamic user engagement through natural dialogue and interpretable suggestions. 3. Simulation-oriented approaches — employ multi-agent frameworks to model complex user-item interactions and system dynamics.

Architectural Foundations

Beyond paradigm categorization, the survey analyzes the essential components of LLM-powered recommendation agents:

Profile construction — building user and item representations
Memory management — retaining and organizing interaction history
Strategic planning — deciding when and how to act (retrieval, tool calls, etc.)
Action execution — carrying out planned steps

The investigation extends to a comprehensive analysis of benchmark datasets and evaluation frameworks in this domain.

Context: Evolution of the Field

The survey situates agentic recommendation within the broader evolution of neural retrieval and generative search:

2019–2021: BERT-based re-ranking and dense retrieval (DPR) established neural retrieval foundations
2022–2023: RAG and retrieval-generation fusion gained traction
2024 onward: Conversational/agentic search and Gen-RecSys exploded
2025–2026: RL-trained search agents, Deep Research, and GraphRAG emerging as growth areas

Four main technical lines are typically compared:

Dense retrieval: high recall, low latency; suited for first-stage retrieval
Late interaction (e.g., ColBERT): higher precision but larger indexes
Generative IR: directly "generates" documents via tokens or docids, simplifying the cascade
Agentic: models search as sequential decision-making, supporting multi-hop reasoning and self-reflection

Key Takeaways for Search/Rec/Personalization

1. Architecture: Cascaded retrieve + rank + generate remains mainstream, but the agentic paradigm makes retrieval count and strategy themselves learnable objects. 2. Data: High-quality instruction data and click/session logs are both critical; synthetic data must guard against leakage and distribution shift. 3. Evaluation: The gap between offline metrics and online satisfaction is widening; LLM-as-judge needs cross-validation with human evaluation. 4. Product: Latency, cost, explainability, and safety are hard constraints for industrial deployment—academic benchmarks alone are insufficient.

Open Challenges & Future Directions

Insufficient unified benchmarks; private data hinders reproducibility
LLM evaluation bias
Safety and cost constraints of agentic systems on the open web
Finer-grained process supervision, retrieval-reasoning co-training, multimodal and cross-lingual consistency
Causal and fairness constraints for recommender systems; deeper integration with knowledge graphs and structured databases

Original Abstract

> Recommender systems are essential components of many online platforms, yet traditional approaches still struggle with understanding complex user preferences and providing explainable recommendations. The emergence of Large Language Model (LLM)-powered agents offers a promising approach by enabling natural language interactions and interpretable reasoning, potentially transforming research in recommender systems. This survey provides a systematic review of the emerging applications of LLM-powered agents in recommender systems. We identify and analyze three key paradigms in current research: (1) Recommender-oriented approaches...; (2) Interaction-oriented approaches...; and (3) Simulation-oriented approaches... Beyond paradigm categorization, we analyze the architectural foundations of LLM-powered recommendation agents, examining their essential components: profile construction, memory management, strategic planning, and action execution. Our investigation extends to a comprehensive analysis of benchmark datasets and evaluation frameworks in this domain.