Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Foundation Models for Recommender Systems: A Survey and New Perspectives

Chengkai Huang, Tong Yu, Kaige Xie, Shuai Zhang, Lina Yao, Julian McAuley

TL;DR

This survey addresses the problem of how to harness Foundation Models to advance recommender systems, focusing on improved generalization, interactive experiences, and explainability. It presents a four-part taxonomy and covers three main FM families—Language Foundation Models, Personalized Agents, and Multi-modal Foundation Models—along with applications in context-aware, interactive, and cross-domain settings. Key contributions include a unified framework that links data representations, model types, and downstream tasks, plus a synthesis of open problems such as long sequences, explainability, efficiency, benchmarks, and safety. The work provides a roadmap showing how retrieval-augmented approaches, knowledge prompts, and agent-based architectures can transform RS performance and user experience, while highlighting practical challenges before FM4RecSys can be deployed at scale.

Abstract

Recently, Foundation Models (FMs), with their extensive knowledge bases and complex architectures, have offered unique opportunities within the realm of recommender systems (RSs). In this paper, we attempt to thoroughly examine FM-based recommendation systems (FM4RecSys). We start by reviewing the research background of FM4RecSys. Then, we provide a systematic taxonomy of existing FM4RecSys research works, which can be divided into four different parts including data characteristics, representation learning, model type, and downstream tasks. Within each part, we review the key recent research developments, outlining the representative models and discussing their characteristics. Moreover, we elaborate on the open problems and opportunities of FM4RecSys aiming to shed light on future research directions in this area. In conclusion, we recap our findings and discuss the emerging trends in this field.

Foundation Models for Recommender Systems: A Survey and New Perspectives

TL;DR

This survey addresses the problem of how to harness Foundation Models to advance recommender systems, focusing on improved generalization, interactive experiences, and explainability. It presents a four-part taxonomy and covers three main FM families—Language Foundation Models, Personalized Agents, and Multi-modal Foundation Models—along with applications in context-aware, interactive, and cross-domain settings. Key contributions include a unified framework that links data representations, model types, and downstream tasks, plus a synthesis of open problems such as long sequences, explainability, efficiency, benchmarks, and safety. The work provides a roadmap showing how retrieval-augmented approaches, knowledge prompts, and agent-based architectures can transform RS performance and user experience, while highlighting practical challenges before FM4RecSys can be deployed at scale.

Abstract

Recently, Foundation Models (FMs), with their extensive knowledge bases and complex architectures, have offered unique opportunities within the realm of recommender systems (RSs). In this paper, we attempt to thoroughly examine FM-based recommendation systems (FM4RecSys). We start by reviewing the research background of FM4RecSys. Then, we provide a systematic taxonomy of existing FM4RecSys research works, which can be divided into four different parts including data characteristics, representation learning, model type, and downstream tasks. Within each part, we review the key recent research developments, outlining the representative models and discussing their characteristics. Moreover, we elaborate on the open problems and opportunities of FM4RecSys aiming to shed light on future research directions in this area. In conclusion, we recap our findings and discuss the emerging trends in this field.
Paper Structure (28 sections, 3 figures, 1 table)

This paper contains 28 sections, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Motivation
  3. Enhanced Generalization Capabilities.
  4. Elevated Recommendation Experience
  5. Improved Explanation & Reasoning Capabilities.
  6. Distinguishing Features from Recent LLM based Recsys Surveys
  7. Differences and Key Contributions:
  8. Research Progress in FM4RecSys
  9. Data Characteristics and Representation Learning
  10. Classification Framework for FM4RecSys
  11. Language Foundation Models for RecSys
  12. Personalized Agent for RecSys with Foundation Models
  13. Multi-modal Foundation Models for RecSys
  14. Applications of FM4RecSys
  15. FM for Context-aware RecSys
  16. ...and 13 more sections

Figures (3)

  • Figure 1: The taxonomy of FM4RecSys from data characteristics to open problems and opportunities.
  • Figure 2: Classification frameworks in FM4RecSys from the Model Type perspective.
  • Figure 3: Two types of personalized agents in FM4RecSys: (a) Agent as User Simulator and (b) Agent as Recommender System.