Flow Matching based Sequential Recommender Model

2025-05-22

AI Summary

Diffusion models suffer from inaccurate user preference modeling in sequential recommendation due to bidirectional noise perturbations in both forward and reverse processes. To address this, we propose FMRec—the first framework to introduce Flow Matching (FM) into sequential recommendation. Our key contributions are: (1) a straight-line flow trajectory design coupled with a task-specific flow matching loss; (2) integration of a forward reconstruction loss to enhance robustness against noise; and (3) replacement of stochastic sampling with a deterministic ODE-based backward generator to improve recommendation consistency. Extensive experiments on four benchmark datasets demonstrate that FMRec consistently outperforms state-of-the-art methods by an average of 6.53% in top-K recommendation metrics. The implementation is publicly available.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Reasoning under Uncertainty: Sequential Decision Making
• Data Mining & Knowledge Management: Recommender Systems

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Generative models, particularly diffusion model, have emerged as powerful tools for sequential recommendation. However, accurately modeling user preferences remains challenging due to the noise perturbations inherent in the forward and reverse processes of diffusion-based methods. Towards this end, this study introduces FMRec, a Flow Matching based model that employs a straight flow trajectory and a modified loss tailored for the recommendation task. Additionally, from the diffusion-model perspective, we integrate a reconstruction loss to improve robustness against noise perturbations, thereby retaining user preferences during the forward process. In the reverse process, we employ a deterministic reverse sampler, specifically an ODE-based updating function, to eliminate unnecessary randomness, thereby ensuring that the generated recommendations closely align with user needs. Extensive evaluations on four benchmark datasets reveal that FMRec achieves an average improvement of 6.53% over state-of-the-art methods. The replication code is available at https://github.com/FengLiu-1/FMRec.

Problem

• Modeling user preferences accurately in sequential recommendation
• Reducing noise perturbations in diffusion-based recommendation methods
• Aligning generated recommendations closely with user needs

Innovation

• Flow Matching model with straight flow trajectory
• Reconstruction loss for noise robustness
• ODE-based deterministic reverse sampler