Image Fusion for Cross-Domain Sequential Recommendation

TL;DR

This paper introduces IFCDSR, a novel cross-domain sequential recommendation method that leverages item image embeddings and attention mechanisms to improve user preference prediction across multiple domains.

Contribution

The paper proposes a new approach integrating visual item data via a frozen CLIP model and a multiple attention layer to enhance cross-domain recommendation accuracy.

Findings

IFCDSR significantly outperforms existing methods on four e-commerce datasets.

Visual information from images improves recommendation quality.

Cross-domain interest modeling benefits from joint learning of intra- and inter-sequence interactions.

Abstract

Cross-Domain Sequential Recommendation (CDSR) aims to predict future user interactions based on historical interactions across multiple domains. The key challenge in CDSR is effectively capturing cross-domain user preferences by fully leveraging both intra-sequence and inter-sequence item interactions. In this paper, we propose a novel method, Image Fusion for Cross-Domain Sequential Recommendation (IFCDSR), which incorporates item image information to better capture visual preferences. Our approach integrates a frozen CLIP model to generate image embeddings, enriching original item embeddings with visual data from both intra-sequence and inter-sequence interactions. Additionally, we employ a multiple attention layer to capture cross-domain interests, enabling joint learning of single-domain and cross-domain user preferences. To validate the effectiveness of IFCDSR, we re-partitioned four e-commerce datasets and conducted extensive experiments. Results demonstrate that IFCDSR significantly outperforms existing methods.