Title: Distinguished Colloquium——Network reconstruction problems with permutational structure
Speaker：Jian Ding (University of Pennsylvania)
Venue：Room 1114, Sciences Building No. 1
Abstract: In this talk I will present three examples on recovering a combinatorial structure of permutational type from noisy data: hidden Hamiltonian cycle problem, random graph matching problem and planted bipartite matching problem. I will emphasize the probabilistic and combinatorial methods in analyzing these statistical models, from both information-theoretic and computational point of view. This is based on joint works with Vivek Bargaria, Zongming Ma, David Tse, Yihong Wu, Jiaming Xu and Dana Yang in various combinations.
Bio: Jian Ding is a Gilbert Helman Professor at University of Pennsylvania. His main research area is in probability theory, with focus on interactions with statistical physics and theoretical computer science. He also has a broad interest in probability questions that arise from "application-oriented" problems. Before joining Penn, he has been a postdoc at Stanford and a faculty at University of Chicago, after his Ph.D. at UC Berkeley in 2011.