Assoc.Prof. Huan Wang

College of Informatics

Huazhong Agricultural University

China

Speech title: A Multi-type Transferable Method for Missing Link Prediction in Heterogeneous Social Networks

Biography

Huan Wang is an associate professor with the College of Informatics, Huazhong Agricultural University, Wuhan 430070, China. His research focuses on social computing, intelligent transportation, and AI for drug discovery and development. Since 2014, he has authored or co-authored over 30 refereed journal and conference papers, such as the IEEE TKDE, ACM TWEB, and ACM TKDD.

Abstract:

Heterogeneous social networks, which are characterized by diverse interaction types, have resulted in new challenges for missing link prediction. Most deep learning models tend to learn type-specific features to maximize the prediction performances on specific link types. However, the types of missing links are uncertain in heterogeneous social networks; this restricts the prediction performance of existing deep learning models. To address this issue, we propose a multi-type transferable method (MTTM) for missing link prediction in heterogeneous social networks, which exploits adversarial neural networks to remain robust against type differences. It comprises a generative predictor and a discriminative classifier. The generative predictor extracts link representations and predicts whether the unobserved link is a missing link. It attempts to capture the shared features among link types to deceive the discriminative classifier to improve the prediction performance. In order not to be deceived, the discriminative classifier attempts to learn type-specific features to accurately distinguish link types. The integrated MTTM is constructed on a minimax two-player game between the generative predictor and the discriminative classifier to predict missing links based on transferable link representations. Extensive experiments show that the proposed   can outperform state-of-the-art baselines for missing link prediction in heterogeneous social networks.