AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Nest Your Adaptive Algorithm for Parameter-Agnostic Nonconvex Minimax Optimization

Junchi YANG, Xiang Li, Niao He

NeurIPS 2022 · Optimization › Nonconvex and smooth optimization · approach: Complexity of nonconvex first-order methods · 0 citations (OpenAlex)

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
Nest the dual maximization in an inner loop with an adaptive stopping criterion instead of simultaneous primal-dual updatesAdaptive minimax optimization · NC-SC minimax problemsconvergence with varying learning-rate ratiosImproved
Use simultaneous non-nested adaptive primal-dual updates without a problem-dependent time-scale ratioNon-nested adaptive methods · Deterministic quadratic exampleconvergence to stationary point: r \leq LHurt
Use simultaneous GDA updates without a problem-dependent time-scale ratioGDA · Deterministic quadratic exampleconvergence to stationary point: r \leq L^2Hurt
Use a learning-rate ratio of 0.01 rather than a larger ratioNon-nested Adam, AMSGrad, and GDA · McCormick test function, stochasticconvergence: 0.01Hurt
Increase the learning-rate ratio to 0.03Non-nested GDA · McCormick test function, stochasticconvergence: 0.01 -> 0.03Improved
Increase the learning-rate ratio to 0.05Non-nested Adam and AMSGrad · McCormick test function, stochasticconvergence: 0.01 -> 0.05Improved
Use NeAda instead of Adam with fixed 15 inner-loop iterationsDRO training · Synthetic datasettest accuracyImproved
Use the same learning rates for primal and dual variables instead of different scalesNeAda · Synthetic DROtest accuracyNo clear effect
Use the same learning rates for primal and dual variables instead of different scalesAdam · Synthetic DROtest accuracy: about 68\%Hurt
Use NeAda instead of Adam with the same learning ratesDRO training · MNISTtest accuracyImproved

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