Department of Mathematics

Mathematical Physics and Gauge Theory

  •  Machine Learning II: Approximation Theory and Quantum Physics
  •  05/04/2017
  •  11:00 AM - 12:00 PM
  •  C304 Wells Hall
  •  Tim Nguyen, MSU

Might the unreasonable effectiveness of deep learning have its origins in the laws of physics? We start by establishing that neural networks are universal approximators, which is closely related to the Kolmogorov-Arnold representation theorem. We then discuss parallels between the hierarchy of layers in deep learning and the hierarchy of scales in the renormalization group analysis of physical systems: both lead to a dramatic reduction in complexity. Time permitting, we discuss the use of restricted Boltzmann machines in solving quantum many body problems.



Department of Mathematics
Michigan State University
619 Red Cedar Road
C212 Wells Hall
East Lansing, MI 48824

Phone: (517) 353-0844
Fax: (517) 432-1562

College of Natural Science