Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Contrastive Framework with User, Item and Review Alignment for Recommendation

Hoang V. Dong, Yuan Fang, Hady W. Lauw

TL;DR

This work tackles sparsity in recommender systems by integrating reviews into learning through a Review-centric Contrastive Alignment Framework for Recommendation (ReCAFR). It introduces two self-supervised contrastive strategies: review-augmented contrastive learning to extract robust user/item signals from reviews, and alignment contrastive learning to unify user, item, and review representations in a shared latent space. Empirical results on multiple benchmarks show that ReCAFR consistently improves over backbones that do not use reviews and over existing review-aware methods, with added robustness to missing reviews and further gains when using LLM-enhanced reviews. The approach offers a practical, modular pathway to leverage textual signals in recommendation while maintaining resilience to incomplete review data and enabling future extensions with richer semantic and graph-based supervision.

Abstract

Learning effective latent representations for users and items is the cornerstone of recommender systems. Traditional approaches rely on user-item interaction data to map users and items into a shared latent space, but the sparsity of interactions often poses challenges. While leveraging user reviews could mitigate this sparsity, existing review-aware recommendation models often exhibit two key limitations. First, they typically rely on reviews as additional features, but reviews are not universal, with many users and items lacking them. Second, such approaches do not integrate reviews into the user-item space, leading to potential divergence or inconsistency among user, item, and review representations. To overcome these limitations, our work introduces a Review-centric Contrastive Alignment Framework for Recommendation (ReCAFR), which incorporates reviews into the core learning process, ensuring alignment among user, item, and review representations within a unified space. Specifically, we leverage two self-supervised contrastive strategies that not only exploit review-based augmentation to alleviate sparsity, but also align the tripartite representations to enhance robustness. Empirical studies on public benchmark datasets demonstrate the effectiveness and robustness of ReCAFR.

A Contrastive Framework with User, Item and Review Alignment for Recommendation

TL;DR

This work tackles sparsity in recommender systems by integrating reviews into learning through a Review-centric Contrastive Alignment Framework for Recommendation (ReCAFR). It introduces two self-supervised contrastive strategies: review-augmented contrastive learning to extract robust user/item signals from reviews, and alignment contrastive learning to unify user, item, and review representations in a shared latent space. Empirical results on multiple benchmarks show that ReCAFR consistently improves over backbones that do not use reviews and over existing review-aware methods, with added robustness to missing reviews and further gains when using LLM-enhanced reviews. The approach offers a practical, modular pathway to leverage textual signals in recommendation while maintaining resilience to incomplete review data and enabling future extensions with richer semantic and graph-based supervision.

Abstract

Learning effective latent representations for users and items is the cornerstone of recommender systems. Traditional approaches rely on user-item interaction data to map users and items into a shared latent space, but the sparsity of interactions often poses challenges. While leveraging user reviews could mitigate this sparsity, existing review-aware recommendation models often exhibit two key limitations. First, they typically rely on reviews as additional features, but reviews are not universal, with many users and items lacking them. Second, such approaches do not integrate reviews into the user-item space, leading to potential divergence or inconsistency among user, item, and review representations. To overcome these limitations, our work introduces a Review-centric Contrastive Alignment Framework for Recommendation (ReCAFR), which incorporates reviews into the core learning process, ensuring alignment among user, item, and review representations within a unified space. Specifically, we leverage two self-supervised contrastive strategies that not only exploit review-based augmentation to alleviate sparsity, but also align the tripartite representations to enhance robustness. Empirical studies on public benchmark datasets demonstrate the effectiveness and robustness of ReCAFR.
Paper Structure (20 sections, 6 equations, 5 figures, 8 tables, 1 algorithm)

This paper contains 20 sections, 6 equations, 5 figures, 8 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Preliminaries
  4. Proposed Approach
  5. Overview
  6. Review-Augmented Contrastive Learning
  7. Alignment Contrastive Strategy
  8. Training and Inference
  9. Experiments
  10. Experimental Setup
  11. Performance Comparison
  12. Ablation Study
  13. Hyperparameter Studies
  14. Can ReCAFR benefit from LLMs?
  15. Conclusion and Future Work
  16. ...and 5 more sections

Figures (5)

  • Figure 1: Comparison of the similarity distribution of the positive pairs versus that of the negative pairs.
  • Figure 2: Overall framework of ReCAFR. For brevity, it illustrates (b) review-augmented contrastive learning and (c) alignment contrastive learning for the user side only; similar procedures are applied to the item side.
  • Figure 3: Impact of missing reviews.
  • Figure 4: Effect of $\lambda_1$ and $\lambda_2$ on ReCAFR.
  • Figure 5: Review enhancement using LLMs, illustrated with the Book dataset.