AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

First-order Convergence Theory for Weakly-Convex-Weakly-Concave Min-max Problems

Mingrui Liu, Hassan Rafique, Qihang Lin, Tianbao Yang

JMLR 2021 · Optimization › Nonconvex and smooth optimization · approach: Gradient descent ascent for minimax

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use extragradient instead of gradient descent as subproblem solverIPP framework · Synthetic min-max, rho=10^-3gradient evaluations: 14857 -> 62Improved
use extragradient instead of gradient descent as subproblem solverIPP framework · Synthetic min-max, rho=10^-7gradient evaluations: >10^8 -> 60Improved
use extragradient instead of gradient descent as subproblem solverIPP framework · Synthetic min-max, rho=10^-1gradient evaluations: 9239 -> 1854Improved
use variance reduction (SVRG) instead of SGD for small target errorIPP framework · Synthetic finite-sum, eps=10^-10, rho=10^-1stochastic gradient evaluations: 12710 -> 1002Improved
use variance reduction (SVRG) instead of SGD for small target errorIPP framework · Synthetic finite-sum, eps=10^-10, rho=10^-3stochastic gradient evaluations: 2515 -> 1503Improved
use variance reduction (SVRG) instead of SGD for loose target errorIPP framework · Synthetic finite-sum, eps=10^-1, rho=10^-1stochastic gradient evaluations: 172 -> 1002Hurt
use variance reduction (SVRG) instead of SGD for loose target errorIPP framework · Synthetic finite-sum, eps=10^-1, rho=10^-3stochastic gradient evaluations: 67 -> 1025Hurt
wrap SGD in the inexact proximal point framework (IPP-SGD)SGD · WGAN/WGAN-GP on CIFAR10 and LSUN Bedroominception scoreImproved
wrap Adam in the inexact proximal point framework (IPP-Adam)Adam · WGAN/WGAN-GP on CIFAR10 and LSUN Bedroominception scoreImproved
use 5 discriminator steps per generator step (ablation)Adam · WGAN traininginception scoreImproved
use 5 discriminator steps per generator step (ablation)SGD or WGAN-GP · WGAN-GP and SGD traininginception scoreNo clear effect
use Adam instead of SGD as inner solver in IPPIPP · GAN trainingconvergence speedImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy