Enhancing Recommender System with Multi-modal Knowledge Graph

Recommender systems have shown great potential to solve the problem of information overload and improve user experience in various online applications. To address the data sparsity and cold start problems in these systems, researchers have proposed leveraging knowledge graphs (KGs) as auxiliary information for recommendations, which contain valuable external knowledge. However, most of these works have neglected the diversity of data types, such as texts and images, in multi-modal knowledge graphs (MMKGs). In this paper, we propose the Multi-modal Knowledge Graph Attention Network to enhance recommender systems by leveraging multi-modal knowledge. Specifically, We propose a multi-modal knowledge graph attention mechanism to facilitate knowledge propagation over MMKGs. The resulting knowledge-enhanced entity embedding representations are utilized for recommendation. Extensive experiments conducted on MovieLens datasets demonstrate the rationality and effectiveness of our model.