Cross-domain Recommender Systems via Multimodal Domain Adaptation

TL;DR

This paper proposes a multimodal domain adaptation method for cross-domain recommender systems that aligns entity embeddings using textual and visual information to address data sparsity.

Contribution

It introduces a novel domain adaptation technique leveraging multi-view representations and domain classifiers for improved entity alignment across domains.

Findings

Effective in reducing data sparsity issues

Improves recommendation accuracy on benchmark datasets

Utilizes multi-view features for robust domain alignment

Abstract

Collaborative Filtering (CF) has emerged as one of the most prominent implementation strategies for building recommender systems. The key idea is to exploit the usage patterns of individuals to generate personalized recommendations. CF techniques, especially for newly launched platforms, often face a critical issue known as the data sparsity problem, which greatly limits their performance. Cross-domain CF alleviates the problem of data sparsity by finding a common set of entities (users or items) across the domains, which then act as a conduit for knowledge transfer. Nevertheless, most real-world datasets are collected from different domains, so they often lack information about anchor points or reference information for entity alignment. This paper introduces a domain adaptation technique to align the embeddings of entities across domains. Our approach first exploits the available textual and visual information to independently learn a multi-view latent representation for each entity in the auxiliary and target domains. The different representations of the entity are then fused to generate the corresponding unified representation. A domain classifier is then trained to learn the embedding for the domain alignment by fixing the unified features as the anchor points. Experiments on \AS{four} publicly available benchmark datasets indicate the effectiveness of our proposed approach.