AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

Probable Domain Generalization via Quantile Risk Minimization

Cian Eastwood, Alexander Robey, Shashank Singh, Julius Von Kügelgen, Hamed Hassani, George J. Pappas, Bernhard Schölkopf

NeurIPS 2022 · Label-efficient and robust learning › Test-time adaptation · approach: Characterizing out-of-distribution failures · 16 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
Optimize an estimated α-quantile of risk across training domains rather than average riskERM-style domain generalization · iWildCam0.90 quantile risk: 2.73 -> 1.81Improved
Optimize an estimated α-quantile of risk across training domains rather than average riskERM-style domain generalization · OGB-MolPCBA1.0 quantile risk: 16.04 -> 7.16Improved
Optimize an estimated α-quantile of risk across training domains rather than average riskERM-style domain generalization · OGB-MolPCBAmean risk: 0.051 -> 0.053Hurt
Vary α from 0.25 to 0.99 (ablation)EQRM · iWildCam0.90 quantile risk: 3.48 -> 1.81Mixed
Vary α from 0.25 to 0.99 (ablation)EQRM · OGB-MolPCBA1.0 quantile risk: 15.46 -> 7.16Mixed
Add ERM pretraining before the DG penalty (ablation)ColoredMNIST DG methods · ColoredMNISTtest accuracyImproved
Increase α toward 1 (ablation)EQRM · ColoredMNISTcross-domain accuracy consistencyImproved
Use KDE smoothing to estimate risk distributions (ablation)EQRM · Domain generalizationtail-risk extrapolationImproved

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