posted on 2024-09-19, 11:19authored byJie He, Lei Zhang, WenMing Cao, Mingming Yang, Man Li, Zihao Zhao, Man-Fai Leung
In this work, we introduce a Multi-modal Bayesian Recommendation System (MBR) that leverages both image and text modalities to enhance recommendation quality using common implicit feedback. Utilizing deep convolutional neural networks for image feature extraction and a language representation model for textual analysis, our proposed method effectively integrates these modalities to understand their collective impact on user preferences. Experiments on a large-scale dataset demonstrate the efficacy of MBR in providing improved recommendations.