AI Paper Atlas

Research map / Learning theory and algorithms / Fairness

Fairness Reprogramming

Guanhua Zhang, Yihua Zhang, Yang Zhang, Wenqi Fan, Qing Li, Sijia Liu, Shiyu Chang

NeurIPS 2022 · Learning theory and algorithms › Fairness · approach: Post-processing and transport for fairness · 6 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
use fairness reprogramming (Border) instead of best post-processing baselineFairReprogram · CelebA EOaccuracy at similar negative EO: 0.951 vs 0.927Improved
use FairReprogram (Soft) instead of best post-processing baselineFairReprogram · Civil Comments EOnegative EO and accuracy: -0.031 -> -0.011Improved
decrease number of trigger words to one (ablation)FairReprogram Soft · Civil Commentsnegative DP bias: -0.002 -> -0.011Hurt
keep a one-word trigger versus no trigger (ablation)FairReprogram · Civil Commentsnegative bias scoreImproved
use five or more trigger words (ablation)FairReprogram · Civil Commentsnegative bias scoreNo clear effect
increase image trigger size (ablation)FairReprogram · CelebAnegative bias score: -0.081 -> -0.054Improved
increase image trigger size (ablation)FairReprogram · CelebAaccuracy: 0.951 -> 0.914Hurt
reduce tuning data to 1 percentFairReprogram · Civil Comments and CelebAnegative bias scoreNo clear effect
reduce tuning data to very smallAdvPost · tuning data rationegative bias scoreHurt
transfer trigger to model trained with different seedFairReprogram · Civil Comments and CelebAfairness-accuracy trade-offNo clear effect
increase number of classes in multi-class settingAdvPost · CelebA multi-classfairness-accuracy trade-offHurt
replace adversarial loss with MMD regularizationMmdIn/MmdPost · Civil Commentsfairness-accuracy trade-offHurt

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