AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Moderately Distributional Exploration for Domain Generalization

Rui Dai, Yonggang Zhang, Zhen Fang, Bo Han, Xinmei Tian

ICML 2023 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Foundation-model domain generalization · 13 citations (OpenAlex)

Read the paper

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
Explore a semantic-preserving non-semantic uncertainty subset with AdaIN-based style transferMODE · PACSaverage accuracy: 86.0 -> 86.9Improved
Explore a semantic-preserving non-semantic uncertainty subset with Fourier-based style transferMODE · PACSaverage accuracy: 86.0 -> 85.7No clear effect
Explore a semantic-preserving non-semantic uncertainty subset with AdaIN-based style transferMODE · Office-Homeaverage accuracy: 66.6 -> 66.9Improved
Explore a semantic-preserving non-semantic uncertainty subset with Fourier-based style transferMODE · Office-Homeaverage accuracy: 66.6 -> 65.4Hurt
Explore a semantic-preserving non-semantic uncertainty subset with Fourier-based style transferMODE · Digits-DGaverage accuracy: 82.5 -> 83.9Improved
Increase the number of inner exploration steps (ablation)MODE · PACS (Art-painting and Sketch target domains)accuracyImproved
Vary the inner exploration steps (ablation)MODE · PACS (Sketch and Art-painting target domains)accuracyImproved
Increase the uncertainty-set size by increasing stylization strength (ablation)MODE · PACS (Art-painting and Sketch target domains)accuracyMixed
Vary the balance between original-data risk and explored-sample risk (ablation)MODE · PACS (Art-painting and Sketch target domains)accuracyMixed
Vary the number of style providers (ablation)MODE · PACS (Art-painting and Sketch target domains)accuracyMixed

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