Pennsylvania State University
Title: What does the calculus of variations and differential equations have to do with data science?
Date: Friday, February 1, 2019
Place and Time: Room 101, Love Building, 3:35-4:25 pm
Refreshments: Room 204, Love Building, 3:00 pm
Abstract. Data science represents some of the biggest opportunities and challenges to science today. However, many of the algorithms underlying machine learning are not well understood. In this talk I will discuss a number of ways that mathematical analysis can help in understanding these algorithms. In particular, I will discuss how a broad range of tools from modern analysis (including differential equations and the calculus of variations) can be used to understand crucial questions in machine learning relating to overfitting, optimization, reinforcement learning, and Bayesian estimation. This represents joint work with a number of statisticians, engineers, and mathematicians, and will seek to be accessible to a broad mathematical audience.