2 datasets found

The proposed Chg2Cap method is constructed by a CNN-based feature extractor that uses the pre-trained ResNet-101 as the backbone, an attentive encoder that consists of a...

Self-supervised dual-channel attentive network for session-based social recommendation.