AI Paper Atlas

Research map / Learning theory and algorithms / Fairness

Noise-tolerant fair classification

Alex Lamy, Ziyuan Zhong, Aditya K Menon, Nakul Verma

NeurIPS 2019 · Learning theory and algorithms › Fairness · approach: Post-processing and transport for fairness · 31 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
Scale the fairness tolerance by the estimated sensitive-attribute noise rateFair classification with corrupted sensitive attributes · COMPASfairness violationImproved
Scale the fairness tolerance instead of using the corrupted sensitive attribute without adjustmentFair classification with corrupted sensitive attributes · COMPAStesting errorImproved
Denoise the sensitive attribute before fair classification rather than scale the toleranceFair classification with corrupted sensitive attributes · COMPAStesting errorHurt
Scale the fairness tolerance by the noise rateFair classification with corrupted sensitive attributes · law schoolaccuracyImproved
Use an estimated noise rate to scale the fairness toleranceFair classification with corrupted sensitive attributes · law schoolfairness violationNo clear effect
Denoise the sensitive attribute rather than scale the fairness toleranceFair classification with corrupted sensitive attributes · law schoolfairness violationHurt
Increase the sensitive-attribute noise rateFair classification with corrupted sensitive attributes · adulttesting errorHurt

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