A0288
Title: Accelerated gradient method for convex-concave saddle-point problems
Authors: Donghwan Kim - KAIST (Korea, South) [presenting]
Abstract: Some recent machine learning problems, such as a generative adversarial network (GAN), require solving large-dimensional saddle-point problems. Gradient descent ascent type methods are widely used to solve such problems. We present some recent progress on accelerating the gradient method for convex-concave saddle-point problems.
