AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Robust Optimization over Multiple Domains

Qi Qian, Shenghuo Zhu, Jiasheng Tang, Rong Jin, Baigui Sun, Hao Li

AAAI 2019 · Optimization › Nonconvex and smooth optimization · approach: Coordinate and Adam convergence analysis · 50 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
Learn model and adversarial domain weights jointly instead of using even weightsMixture_Opt versus Mixture_Even · Pets categorizationworst-case test accuracy (%): 88.50 -> 89.35Improved
Learn model and adversarial domain weights jointly instead of using even weightsMixture_Opt versus Mixture_Even · Digits recognitionworst-case test accuracy (%): 91.33 -> 92.14Improved
Train one model on multiple domains rather than an individual domainPets categorization · Pets categorizationworst-case test accuracy (%): 80.44 -> 89.35Improved
Train one model on multiple domains rather than an individual domainDigits recognition · Digits recognitionworst-case test accuracy (%): 68.25 -> 92.14Improved
Use a strongly concave L2 regularizer instead of no distribution regularizer (ablation)Robust optimization over multiple domains · Noisy ImageNet pets domainsworst-case performanceImproved
Constrain adversarial weights toward a prior using the regularization strength (ablation)Mixture_Opt · Pets categorizationinter-domain loss discrepancyImproved
Use the OT distribution regularizer instead of the L2 regularizer (ablation)Mixture_Opt · Pets categorizationinter-domain loss discrepancyMixed
Use the delayed distribution step size 1/(lambda(t+c)) with an appropriate c (ablation)Regularized robust optimization · Non-convex convergence analysisadversarial-distribution regretImproved

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