Uncertainty-Aware Pseudo-Labeling and Dual Graph Driven Network for Incomplete Multi-View Multi-Label Classification

Wulin Xie, Xiaohuan Lu, Yadong Liu, Jiang Long, Bob Zhang, Shuping Zhao, Jie Wen

Abstract: Multi-view multi-label classification has recently received extensive attention due to its wide-ranging applications across various fields, such as medical imaging and bioinformatics. However, views and labels are usually incomplete in practical scenarios, attributed to the uncertainties in data collection and manual labeling. To cope with this issue, we propose an uncertainty-aware pseudo-labeling and dual graph driven network (UPDGD-Net), which can fully leverage the supervised information of the available labels and feature information of available views. Different from the existing works, we leverage the label matrix to impose dual graph constraints on the embedded features of both view-level and label-level, which enables the method to maintain the inherent structure of the real data during the feature extraction stage. Furthermore, our network incorporates an uncertainty-aware pseudo-labeling strategy to fill the missing labels, which not only addresses the learning issue of incomplete multi-labels but also enables the method to explore more supervised information to guide the network training. Extensive experiments on five datasets demonstrate that our method outperforms other state-of-the-art methods.

Primary Subject Area: [Generation] Multimedia Foundation Models

Secondary Subject Area: [Experience] Multimedia Applications

Relevance To Conference: This work introduces the Uncertainty-Aware Pseudo-Labeling and Dual Graph Driven Network (UPDGD-Net), a novel framework specifically designed to address the challenges of incomplete data in multi-view multi-label classification, a prevalent issue in multimedia/multimodal processing. By leveraging an uncertainty-aware pseudo-labeling strategy and a label-guided dual-constraints approach, UPDGD-Net effectively utilizes partial labels and views to maintain the integrity and richness of multi-label information. This approach enhances the robustness and accuracy of classifications across various domains, including those rich in multimedia content like medical imaging and bioinformatics.

Submission Number: 954