https://www.math.snu.ac.kr/board/files/attach/images/701/ff97c54e6e21a4ae39315f9a12b27314.png
Extra Form
강연자 이창한
소속 Northwestern University
date 2021-09-16

 

Abstract: 
While the typical behaviors of stochastic systems are often deceptively oblivious to the tail distributions of the underlying uncertainties, the ways rare events arise are vastly different depending on whether the underlying tail distributions are light-tailed or heavy-tailed. Roughly speaking, in light-tailed settings, a system-wide rare event arises because everything goes wrong a little bit as if the entire system has conspired up to provoke the rare event (conspiracy principle), whereas, in heavy-tailed settings, a system-wide rare event arises because a small number of components fail catastrophically (catastrophe principle). In the first part of this talk, I will introduce the recent developments in the theory of large deviations for heavy-tailed stochastic processes at the sample path level and rigorously characterize the catastrophe principle. In the second part, I will explore an intriguing connection between the catastrophe principle and a central mystery of modern AI—the unreasonably good generalization performance of deep neural networks.
 
This talk is based on the ongoing research in collaboration with Mihail Bazhba, Jose Blanchet, Bohan Chen, Sewoong Oh, Insuk Seo, Zhe Su, Xingyu Wang, and Bert Zwart.
 
Short Bio: 
Chang-Han Rhee is an Assistant Professor in Industrial Engineering and Management Sciences at Northwestern University. Before joining Northwestern University, he was a postdoctoral researcher in the Stochastics Group at Centrum Wiskunde & Informatica and in Industrial & Systems Engineering and Biomedical Engineering at Georgia Tech. He received his Ph.D. in Computational and Mathematical Engineering from Stanford University. His research interests include applied probability, stochastic simulation, and statistical learning. He was a winner of the Outstanding Publication Award from the INFORMS Simulation Society in 2016, a winner of the Best Student Paper Award (MS/OR focused) at the 2012 Winter Simulation Conference, and a finalist of the 2013 INFORMS George Nicholson Student Paper Competition.
Atachment
첨부 '1'
List of Articles
카테고리 제목 소속 강연자
수학강연회 Green’s function for initial-boundary value problem file National Univ. of Singapore Shih-Hsien Yu
수학강연회 Heavy-tailed large deviations and deep learning's generalization mystery file Northwestern University 이창한
수학강연회 Solver friendly finite element methods file Oklahoma State Univ. 구자언
수학강연회 Deformation spaces of Kleinian groups and beyond file Osaka University Kenichi Ohshika
수학강연회 On some nonlinear elliptic problems file Paul Sabatier University, Toulouse Yuri Egorov
수학강연회 Partial differential equations with applications to biology file POSTECH 황형주
수학강연회 Limit computations in algebraic geometry and their complexity file POSTECH 현동훈
수학강연회 Variational Methods without Nondegeneracy file POSTECH 변재형
수학강연회 Compressible viscous Navier-Stokes flows: Corner singularity, regularity file POSTECH 권재룡
수학강연회 Mixed type PDEs and compressible flow file POSTECH 배명진
수학강연회 Connes's Embedding Conjecture and its equivalent file RIMS Narutaka Ozawa
수학강연회 Recent progress on the Brascamp-Lieb inequality and applications file Saitama University Neal Bez
수학강연회 A-infinity functor and topological field theory file Simons Center for Geometry and Physics Kenji Fukaya
수학강연회 The Shape of Data file Stanford University Gunnar E. Carlsson
수학강연회 On the resolution of the Gibbs phenomenon file SUNY Buffalo 정재훈
수학강연회 <학부생을 위한 ε 강연> What mathematics can do for the real and even fake world file UCLA Stanley Osher
수학강연회 학부생을 위한 강연: 건축과 수학 file UI 건축사무소 위진복
수학강연회 <학부생을 위한 ɛ 강연> Introduction to the incompressible Navier-Stokes equations file UNIST 배한택
수학강연회 Quantitative residual non-vanishing of special values of various L-functions file UNIST 선해상
수학강연회 Counting number fields and its applications file UNIST 조재현
Board Pagination Prev 1 2 3 4 5 6 7 8 9 10 11 12 Next
/ 12