Personalized Diffusion Model Reshapes Cold-Start Bundle Recommendation

2025-05-20

AI Summary

To address the failure of collaborative filtering in cold-start bundle recommendation—caused by extreme sparsity in user-bundle interactions—this paper proposes DisCo, a generative bundle recommendation framework. DisCo’s key innovations are: (1) the first integration of disentangled, personalized user interest representations into a diffusion model, mitigating representation degradation under cold-start conditions; and (2) a debiased Top-K recommendation objective loss, specifically designed to alleviate distribution mismatch and bias accumulation inherent in generative models for recommendation. Extensive experiments on three real-world datasets demonstrate that DisCo consistently outperforms five state-of-the-art baselines across accuracy metrics (HR@K, NDCG@K) and diversity (ILS), confirming the effectiveness and robustness of generative modeling for cold-start bundle recommendation.

Technology Category

• Data Mining & Knowledge Management: Recommender Systems
• Machine Learning: Learning Preferences or Rankings
• Search and Optimization: Mixed Discrete/Continuous Search

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: Graph embeddings and representation learning for Web-related graphs

Abstract

Bundle recommendation aims to recommend a set of items to each user. However, the sparser interactions between users and bundles raise a big challenge, especially in cold-start scenarios. Traditional collaborative filtering methods do not work well for this kind of problem because these models rely on interactions to update the latent embedding, which is hard to work in a cold-start setting. We propose a new approach (DisCo), which relies on a personalized Diffusion backbone, enhanced by disentangled aspects for the user's interest, to generate a bundle in distribution space for each user to tackle the cold-start challenge. During the training phase, DisCo adjusts an additional objective loss term to avoid bias, a prevalent issue while using the generative model for top-$K$ recommendation purposes. Our empirical experiments show that DisCo outperforms five comparative baselines by a large margin on three real-world datasets. Thereby, this study devises a promising framework and essential viewpoints in cold-start recommendation. Our materials for reproducibility are available at: https://github.com/bt-nghia/DisCo.

Problem

• Addresses cold-start bundle recommendation challenges
• Overcomes sparse user-bundle interaction limitations
• Introduces personalized Diffusion model for unbiased recommendations

Innovation

• Personalized Diffusion backbone for bundle generation
• Disentangled aspects enhance user interest modeling
• Additional objective loss term reduces bias