A Survey on Generative Recommendation: Data, Model, and Tasks

2025-10-31

AI Summary

Traditional discriminative recommender systems suffer from limitations in semantic understanding, multi-step reasoning, and interactive capability. Method: This work reframes recommendation as a generative paradigm, proposing a three-tier technical framework—data augmentation, model alignment, and task execution—that integrates large language models (LLMs) and diffusion models, augmented with knowledge injection, agent simulation, heterogeneous signal unification, and prompt optimization. Contributions: (1) We introduce the first three-dimensional analytical framework that systematically characterizes five key advantages of generative recommendation: world-knowledge integration, multi-step reasoning, creative content generation, cross-domain generalization, and natural interaction. (2) We comprehensively survey data construction paradigms, model taxonomies, and task formulations for generative recommendation. (3) We delineate the technical roadmap and future directions toward intelligent recommendation assistants endowed with understanding, reasoning, and generative capabilities.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Recommender Systems

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Recommender systems serve as foundational infrastructure in modern information ecosystems, helping users navigate digital content and discover items aligned with their preferences. At their core, recommender systems address a fundamental problem: matching users with items. Over the past decades, the field has experienced successive paradigm shifts, from collaborative filtering and matrix factorization in the machine learning era to neural architectures in the deep learning era. Recently, the emergence of generative models, especially large language models (LLMs) and diffusion models, have sparked a new paradigm: generative recommendation, which reconceptualizes recommendation as a generation task rather than discriminative scoring. This survey provides a comprehensive examination through a unified tripartite framework spanning data, model, and task dimensions. Rather than simply categorizing works, we systematically decompose approaches into operational stages-data augmentation and unification, model alignment and training, task formulation and execution. At the data level, generative models enable knowledge-infused augmentation and agent-based simulation while unifying heterogeneous signals. At the model level, we taxonomize LLM-based methods, large recommendation models, and diffusion approaches, analyzing their alignment mechanisms and innovations. At the task level, we illuminate new capabilities including conversational interaction, explainable reasoning, and personalized content generation. We identify five key advantages: world knowledge integration, natural language understanding, reasoning capabilities, scaling laws, and creative generation. We critically examine challenges in benchmark design, model robustness, and deployment efficiency, while charting a roadmap toward intelligent recommendation assistants that fundamentally reshape human-information interaction.

Problem

• Surveying generative recommendation systems using data, model, and task frameworks
• Reconceptualizing recommendation as generation rather than discriminative scoring
• Examining generative models' capabilities in knowledge integration and personalized content

Innovation

• Generative models reconceptualize recommendation as generation task
• Unified framework spans data augmentation and model alignment
• LLMs enable conversational interaction and personalized content generation