Review-Based Domain Disentanglement without Duplicate Users or Contexts for Cross-Domain Recommendation

Abstract

A cross-domain recommendation has shown promising results in solving data-sparsity and cold-start problems. Despite such progress, existing methods focus on domain-shareable information (overlapped users or same contexts) for a knowledge transfer, and they fail to generalize well without such requirements. To deal with these problems, we suggest utilizing review texts that are general to most e-commerce systems. Our model (named SER) uses three text analysis modules, guided by a single domain discriminator for disentangled representation learning. Here, we suggest a novel optimization strategy that can enhance the quality of domain disentanglement, and also debilitates detrimental information of a source domain. Also, we extend the encoding network from a single to multiple domains, which has proven to be powerful for review-based recommender systems. Extensive experiments and ablation studies demonstrate that our method is efficient, robust, and scalable compared to the state-of-the-art single and cross-domain recommendation methods.

Proceedings of the 31st ACM International Conference on Information & Knowledge Management

References in corpus (15)

• Domain Separation Networks
• Maximum Density Divergence for Domain Adaptation
• Neural Rating Regression with Abstractive Tips Generation for Recommendation
• CATN: Cross-Domain Recommendation for Cold-Start Users via Aspect Transfer Network
• A Review-aware Graph Contrastive Learning Framework for Recommendation
• Domain Agnostic Learning with Disentangled Representations
• Learning Disentangled Semantic Representation for Domain Adaptation
• Transfer Learning via Contextual Invariants for One-to-Many Cross-Domain Recommendation
• How Useful are Reviews for Recommendation? A Critical Review and Potential Improvements
• Cross-Market Product Recommendation
• DDTCDR: Deep Dual Transfer Cross Domain Recommendation
• Joint Contrastive Learning for Unsupervised Domain Adaptation
• AutoFT: Automatic Fine-Tune for Parameters Transfer Learning in Click-Through Rate Prediction
• A Zero Attentive Relevance Matching Networkfor Review Modeling in Recommendation System
• Domain Adaptation via Maximizing Surrogate Mutual Information

Cited by in corpus (2)

• Domain Disentanglement with Interpolative Data Augmentation for Dual-Target Cross-Domain Recommendation
• Review-Based Hyperbolic Cross-Domain Recommendation