MMGCF: Multimodal Graph Collaborative Filtering for Recommendation with Graph Convolutional Networks

The growing interest of the research community in multimodal recommendation models has introduced significant challenges, most notably how to effectively process and integrate multimodal features into traditional approaches such as collaborative filtering. Despite the efforts, many existing methods achieve only marginal performance gains at the cost of increased model complexity and longer training time. In this paper, we present MMGCF (MultiModal Graph Collaborative Filtering), a recommendation model that integrates an efficient collaborative filtering backbone with pre-trained multimodal item features, thereby combining the advantages of multimodal representations with negligible additional model complexity. Through extensive experiments, we evaluate MMGCF against state-of-the-art recommendation models. The results demonstrate that our model achieves significant performance gains with minimum overhead, maintaining competitive training times. Our source code is available at https://github.com/swapUniba/MMGCF