Academic Report of School of Mathematical Sciences [2026] No. 087

(Series Report for High-Level University Construction No. 1346)

Title：Influence Maximization via Information Cascade Embedding

Speaker：Professor Wang Junhui(The Chinese University of Hong Kong)

Time：16:40-17:30, Sept. 12, 2026

Location：Room 601, Huiyan Building, Yuehai Campus

Abstract: Identifying influential users from information cascades is a central task in network science, yet it is often complicated by unobserved network structure. In this talk, we ppropose a cascade embedding model that embeds each user into a latent space and treats the diffusion process as a self-activated survival model without requiring an observed network. Particularly, the hazard function is specified via the heat kernel form, capturing the physical phenomenon of heat diffusion’s rise-to-peak and decay dynamics. The proposed model is identifiable up to rigid motions of the embedding, including translation, rotation, and reflection. Asymptotic properties are established based on the geometry of the quotient space through maximizing likelihood.

Building upon the fitted embedding, we formulate the classical multiple-seed influence maximization problem, establish monotonicity and submodularity of the expected spread, and quantify how embedding estimation and Monte Carlo errors affect greedy seed selection. Its effectiveness is further demonstrated through extensive simulations and an application to meme diffusion across websites.

Speaker Profile：Professor Wang Junhui is currently a professor and chair of the Department of Statistics at The Chinese University of Hong Kong. He received his bachelor’s degree from Peking University and his Ph.D. in statistics from the University of Minnesota. He is an IMS Fellow and an Elected Member of the International Statistical Institute (ISI). His research interests include statistical machine learning and its applications in biomedicine, economics and finance, and information technology. His research findings have been widely published in top-tier journals and conferences in statistics and machine learning, such as JASA, Biometrika, JMLR, and NeurIPS, and he serves as an associate editor for leading journals including JASA, AoAS, JCGS, and Statistica Sinica.

Faculty and students are welcome to attend!

Invited by: Zhou Yan

School of Mathematical Sciences

September 9, 2026