Murat A. Erdogdu

University of Toronto · Department of Computer Science · Department of Statistics Vector Institute Contact Pratt 286b, 6 King's College Rd. Toronto, ON M5S 3G4 erdogdu@cs.toronto.edu

I am an assistant professor at the University of Toronto in the Department of Computer Science and the Department of Statistics. I am a faculty member of the Machine Learning Group and the Vector Institute, and a CIFAR Chair in Artificial Intelligence. Before, I was a postdoctoral researcher at Microsoft Research - New England. I did my Ph.D. in the Department of Statistics at Stanford University. I have an M.Sc. degree in Computer Science from Stanford, and B.Sc. degrees in Electrical Engineering and Mathematics, both from Bogazici University.

Research Interests

• Optimization and Sampling

• Machine Learning Theory

• Stein's Method

• High-dimensional Statistics

Recent Papers

• A. Mousavi-Hosseini, S. Park, M. Girotti, I. Mitliagkas,  M.A. Erdogdu, Neural Networks Efficiently Learn Low-Dimensional Representations with SGD,  2022

• S. Park,  U. Simsekli,  and  M.A. Erdogdu,   Generalization  Bounds  for  Stochastic Gradient Descent via Localized \(\epsilon\)-Covers,  2022

• J. Ba, M.A. Erdogdu, T. Suzuki, Z. Wang, D. Wu, G. Yang, High-dimensional Asymptotics of Feature Learning: How One Gradient Step Improves the Representation,  2022

• N.M. Vural, L. Yu, K. Balasubramanian, S. Volgushev,  M.A. Erdogdu, Mirror Descent Strikes Again: Optimal Stochastic Convex Optimization under Infinite Noise Variance,  2022

• S. Chewi, M.A. Erdogdu, M. Li, R. Shen, M.S. Zhang,  Analysis of Langevin Monte Carlo from Poincaré to Log-Sobolev,  2022