Mar 8, 2024 Friday
Abstract:
With the explosive growth of online information, recommender systems have played a vital role in addressing the information overload problem. Traditional recommendation frameworks designed methods to improve the recommendation accuracy. However, recent challenges of reliable recommender system has attracted more attentions, especially when facing the increasing requirements of reliable and trustworthy AI applications. In this talk, we study knowledge intelligence driven reliable recommendation algorithms, and concentrate on the explainability and robustness of recommender system. First, we investigate the interpretability of recommendation methods, and propose Knowledge Enhanced Graph Neural Networks (KEGNN) for explainable recommendation by leveraging common-sense knowledge to enhance the representation learning of user behaviors and devising a graph neural networks based user behavior learning and reasoning model for comprehensive knowledge reasoning. KEGNN can provide accurate recommendation results with human-like semantic explanations. Second, we leverage the model co-training knowledge to address the extremely serious issues duo to noisy user feedback. A semi-supervised co-training algorithm is devised for denoising learning (SSCDL) in recommender system and makes robust recommendation. SSCDL algorithm has achieved superior performance in both effectiveness and robustness, and had practical usage in DiDi intelligent transportation sector.
Time:
Mar 8, 2024 Friday
11:00-11:50
Location:
Rm W1-101, GZ Campus
Online Zoom
Join Zoom at: https://hkust-gz-edu-cn.zoom.us/j/4236852791 OR 423 685 2791
Speaker Bio:
Ziyu LYU
Associate Professor
School of Cyber Science and Technology, Sun Yat-sen University
Ziyu Lyu is currently an Associate Professor at School of Cyber Science and Technology, Sun Yat-sen University. Before joining SYSU, she was an Associate Professor in Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences. She received her Ph.D. degree from the Department of Computer Science, The University of Hong Kong in 2016, supervised by Prof. Nikos Mamoulis and Prof. David Wai-lok Cheung. Her research interests include secure and trusted AI, recommender system, nature language processing, and Knowledge Intelligence Powered Applications.