AI Paper Atlas

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

Domain Adaptation meets Individual Fairness. And they get along.

Debarghya Mukherjee, Felix Petersen, Mikhail Yurochkin, Yuekai Sun

NeurIPS 2022 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Statistical alignment and hypothesis testing

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
apply SenSeI individual fairness regularizerERM baseline · ToxicityTNR (annotations): 79.4% ± 2.2% -> 83.5% ± 1.7%Improved
apply SenSeI individual fairness regularizerERM baseline · Biosworst per-gender accuracy: 77.9% ± 0.4% -> 80.2% ± 0.4%Improved
apply SenSeI individual fairness regularizerERM baseline · ToxicityTNR (identity tokens): 75.0% ± 2.3% -> 79.4% ± 1.5%Improved
apply SenSRERM baseline · ToxicityTNR (annotations): 79.4% ± 2.2% -> 81.5% ± 1.1%Improved
apply CLP counterfactual augmentationERM baseline · Biosworst per-gender accuracy: 77.9% ± 0.4% -> 79.9% ± 0.3%Improved
apply GLIF using unlabeled target samplesERM baseline · Toxicitybalanced accuracy: 80.7% ± 0.2% -> 70.5% ± 7.1%Hurt
apply GLIF using unlabeled target samplesERM baseline · ToxicityTNR (annotations): 79.4% ± 2.2% -> 87.0% ± 9.8%Improved
apply GLIFERM baseline · Biosworst per-gender accuracy: 77.9% ± 0.4% -> 77.6% ± 1.0%No clear effect
apply WDA domain adaptation on fairness-motivated data splitERM baseline · Toxicityprediction consistency: 62.1% ± 1.4% -> 65.4% ± 1.3%Improved
apply WDA domain adaptation on fairness-motivated data splitERM baseline · Biosprediction consistency: 94.2% ± 0.1% -> 95.5% ± 0.3%Improved
apply WDAERM baseline · Biosbalanced accuracy: 84.2% ± 0.2% -> 83.3% ± 0.3%Hurt
apply DANNERM baseline · Toxicityprediction consistency: 62.1% ± 1.4% -> 62.8% ± 1.1%No clear effect

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