Graph Foundation Models for Recommendation: A Comprehensive Survey
Source: https://arxiv.org/abs/2502.08346 | Published: 2025-02-12 | Type: Survey
Overview
Recommender systems (RS) serve as a fundamental tool for navigating the vast expanse of online information, with deep learning advancements playing an increasingly important role in improving ranking accuracy. Among these, graph neural networks (GNNs) excel at extracting higher-order structural information, while large language models (LLMs) are designed to process and comprehend natural language, making both approaches highly effective and widely adopted.
Recent research has focused on graph foundation models (GFMs), which integrate the strengths of GNNs and LLMs to model complex RS problems more efficiently by leveraging the graph-based structure of user-item relationships alongside textual understanding.
Key Contributions
- Provides a comprehensive overview of GFM-based recommender system technologies
- Introduces a clear taxonomy of current approaches
- Dives into methodological details of representative techniques
- Highlights key challenges and future directions
- Synthesizes recent advancements to offer insights into the evolving GFM-for-recsys landscape
- 2019-2021: BERT reranking and dense retrieval (DPR) establish neural retrieval foundations
- 2022-2023: RAG and retrieval-generation fusion gain traction
- 2024 onward: Conversational/agentic search and Gen-RecSys proliferate
- 2025+: RL-trained search agents, Deep Research, and GraphRAG emerge as growth areas
- Dense retrieval: high recall, low latency, suited to first-stage retrieval
- Late interaction (e.g., ColBERT): higher accuracy but larger indexes
- Generative IR: directly "generates" documents via tokens/docids, simplifying cascades
- Agentic search: frames search as sequential decision-making supporting multi-hop reasoning and self-reflection
- A Comprehensive Survey on Cross-Domain Recommendation
- A Comprehensive Survey on Retrieval Methods in Recommender Systems
- A Survey of Generative Search and Recommendation in the Era of Large Language Models
- A Survey on LLM-powered Agents for Recommender Systems
Context within the Field
The survey situates GFMs within the broader evolution of neural search and recommendation:
Common methodological threads compared in this space include:
Challenges and Open Problems
1. Architecture: Cascaded retrieval + rerank + generation remains mainstream, but agentic paradigms are making retrieval strategy itself a learnable object 2. Data: High-quality instruction data and click/session logs are critical; synthetic data must guard against knowledge 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. Deployment: Latency, cost, explainability, and safety are hard constraints for industrial adoption
Future directions include unified benchmarks, finer-grained process supervision, retrieval-reasoning co-training, multimodal and cross-lingual consistency, and causal/fairness constraints for recommenders.
Original Abstract (Excerpt)
> Recommender systems (RS) serve as a fundamental tool for navigating the vast expanse of online information... Recent research has focused on graph foundation models (GFMs), which integrate the strengths of GNNs and LLMs to model complex RS problems more efficiently by leveraging the graph-based structure of user-item relationships alongside textual understanding. In this survey, we provide a comprehensive overview of GFM-based RS technologies by introducing a clear taxonomy of current approaches, diving into methodological details, and highlighting key challenges and future directions.