Abstract
Recommendation systems have become an important solution to informationsearch problems. This article proposes a neural matrix factorizationrecommendation system model based on the multimodal large language model calledBoNMF. This model combines BoBERTa's powerful capabilities in natural languageprocessing, ViT in computer in vision, and neural matrix decompositiontechnology. By capturing the potential characteristics of users and items, andafter interacting with a low-dimensional matrix composed of user and item IDs,the neural network outputs the results. recommend. Cold start and ablationexperimental results show that the BoNMF model exhibits excellent performanceon large public data sets and significantly improves the accuracy ofrecommendations.