Abstract:
Machine learning seeks to design a machine that can learn from data. In this 
area, there has been great advancements in algorithms, theories, and 
applications over the past few decades. In this talk, I will introduce the 
mathematical problems underlying different branches of machine learning, e.g., 
supervised learning, unsupervised learning, and reinforcement learning. In each 
of these branches, important mathematical results will be presented. Open 
problems as well as their difficulties will be discussed. I will also provide 
examples to illustrate the application of machine learning methods.

(Following Sanjeev Arora's paper "Mathematics of Machine Learning: An introduction", Proceedings ICM 2018)