Some mathematical aspects of machine learning
[Math. Dept.]
December 10, 2018 10:30-11:30
W303 School of Mathematics
SPEAKER
马志明(中国科学院数学与系统科学研究院)
ABSTRACT
I shall briefly report some progress on machine learning, concerning the Optimizing Neural Networks in the Equivalence Class Space, Differential Equations for Asynchronous Optimization Algorithms, and the Convergence Rate of Asynchronous Stochastic Gradient Descent.
The progress is being made by my students and the colleagues working in MSRA, and its developement involves some mathematical thinking and mathematical methods.
SUPPORTED BY
School of Mathematics, Sichuan University