Shoujin of University of Technology Sydney

Abstract: 

Recent advances in generative AI, large language models (LLMs), and multimodal models are reshaping the landscape of recommender systems and intelligent decision support. While these technologies offer unprecedented capabilities for modeling complex user preferences, generating personalized content, and understanding multimodal information, they also raise critical challenges related to credibility, fairness, explainability, safety, and causal reasoning.

In this talk, I will present a series of recent research advances that collectively move recommender systems beyond accuracy-driven optimization toward trustworthy and human-centered AI. First, I will introduce credibility-aware generative recommendation frameworks based on diffusion models, demonstrating how generation processes can be steered toward more credible and reliable content recommendations. I will then discuss fairness-aware generative recommendation, including a modality-diffused counterfactual framework that addresses missing-modal scenarios while mitigating bias, and a retraining-free approach that promotes fairness in LLM-based recommender systems with minimal deployment cost. Then, I will present a risk-aware reasoning framework for explainable and safe medication recommendation, illustrating how generative and reasoning-based AI can support high-stakes decision making in healthcare settings. Building upon recommendation foundations, I will briefly explore how large language models can uncover causal relationships hidden in multimodal data, providing a new pathway toward interpretable and causally grounded AI systems. The talk will conclude with a discussion of emerging opportunities and open challenges for building next-generation recommendation and decision-support systems that responsibly assist human decision making across diverse domains.

Biography:

Shoujin Wang has been a Lecturer in Data Science at the University of Technology Sydney since 2022. He obtained his PhD in Data Science from the University of Technology Sydney in 2019. He has been continuously named in Stanford’s List of World’s Top 2% Scientists since 2023. His main research interests include data science, machine learning, recommender systems, and trustworthy AI. He has published more than 100 research papers in these areas, most of which appeared in premier data science and AI conferences or journals, including NeurIPS, ICLR, KDD, The WebConf, SIGIR, AAAI, IJCAI, TKDE, TOIS, and CSUR. His research has been broadly reported by national media outlets, including SBS and ABC Science Show, and cited in policy and government-related documents, including European platforms for national news accessible in all EU languages. Shoujin has actively served the research community in various roles, such as Local Co-Chair of PAKDD 2025, Lead Social Chair of NeurIPS 2026, Senior Program Committee member for KDD, IJCAI, and AAAI, and Associate Editor for ACM Transactions on Recommender Systems. He is the recipient of multiple prestigious awards and honours, including the 2026 Young Investigator Award from Applied Sciences, the AAAI 2026 Outstanding Senior Program Committee Service Award, the 2024 NSW iAwards, the 2023 Royal Society of New South Wales Bicentennial Early Career Research and Service Citations Award, and the 2022 IEEE DSAA Next-Generation Data Scientist Award.

Email:

Address:Data Science Institute, University of Technology Sydney, 15 Broadway, Sydney, New South Wales, Australia, 2007