Multi-modal Algorithms as a Promising Approach to Improving Recommendation Accuracy by Leveraging Additional Sources of Information

Abstract

This paper explores the effect of multi-modal algorithms in recommendation systems. The objective is to identify and analyze the methods employed in recent research to improve recommendation accuracy by leveraging additional sources of information. The findings suggest that multi-modal algorithms are a promising approach to improving recommendation accuracy, particularly when combining explicit and implicit feedback, incorporating context awareness, and leveraging multiple types of data sources. The novelty of this approach lies in its ability to incorporate different modalities of information, such as text, images, and user behavior, to enhance the accuracy of the recommendation system and incorporate various techniques such as context-aware user-item embedding, cross-modality utilization, and multimodal embedding fusion-based recommendation. This review provides valuable insights into the effectiveness of multi-modal algorithms in improving recommendation accuracy and can guide future research in this area.