In recommender systems. the lack of interaction data between users and items tends to lead to the problem of data sparsity and cold starts. Recently. the interest modeling frameworks incorporating multi-modal features are widely used in recommendation algorithms. https://www.scrapajul.com/product-category/vitamins/