Applied Maths Seminar: Tam -- Splitting algorithms for training GANs

Dear all, 

Our upcoming AM seminar is held next Wednesday 20 September at 1pm in F07 Carslaw
Lecture Theatre 373.  Our speaker is Matthew Tam (University of Melbourne).  

Talk details follow below: 

--------------------------------------------------------------------------------------------------------------------------------------------
Title: Splitting algorithms for training GANs 

Abstract: Generative adversarial networks (GANs) are an approach to fitting generative
models over complex structured spaces.  Within this framework, the fitting problem is
posed as a zero-sum game between two competing neural networks which are trained
simultaneously.  Mathematically, this problem takes the form of a saddle-point problem;
a well-known example of the type of problem where the usual (stochastic) gradient
descent-type approaches used for training neural networks fail.  In this talk, we
rectify this shortcoming by proposing a new method for training GANs that has: (i) a
sounds theoretical foundation, and (ii) does not increase the algorithm's per iteration
complexity (as compared to gradient descent).  The theoretical analysis is performed
within the framework of monotone operator splitting.
-------------------------------------------------------------------------------------------------------------------------------------------- 

An ongoing list of AM seminars is posted here:
https://www.maths.usyd.edu.au/u/SemConf/Applied.html 

See you there, Jae Min