Abstract: Multimodal embeddings of items and users, along with the loss function in the prediction model, are crucial for multimodal recommendation. Previous studies focused on feature extraction and ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results