AI Paper Atlas

Research map / Learning theory and algorithms / Fairness

Post-processing for Individual Fairness

Felix Petersen, Debarghya Mukherjee, Yuekai Sun, Mikhail Yurochkin

NeurIPS 2021 · Learning theory and algorithms › Fairness · approach: Fairness-constrained classification · 3 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
apply GLIF post-processing to BERT predictionsBERT on Bios · BiosPred. Consist.: 0.942 ± 0.002 -> 0.986 ± 0.002Improved
apply GLIF post-processing to BERT predictionsBERT on Bios · BiosTest Acc.: 0.846 ± 0.003 -> 0.830 ± 0.004Hurt
replace the unnormalized Laplacian with the normalized random walk Laplacian (ablation)GLIF on Bios · BiosPred. Consist.: 0.986 ± 0.002 -> 0.988 ± 0.002Improved
replace the unnormalized Laplacian with the normalized random walk Laplacian (ablation)GLIF on Bios · BiosTest Acc.: 0.830 ± 0.004 -> 0.834 ± 0.003Improved
apply GLIF post-processing to BERT predictionsBERT on Toxicity · ToxicityPred. Consist.: 0.614 ± 0.013 -> 0.835 ± 0.012Improved
apply GLIF post-processing to BERT predictionsBERT on Toxicity · ToxicityTest Acc.: 0.809 ± 0.004 -> 0.803 ± 0.003Hurt
replace the unnormalized Laplacian with the normalized random walk Laplacian (ablation)GLIF on Toxicity · ToxicityPred. Consist.: 0.835 ± 0.012 -> 0.844 ± 0.013Improved
replace the unnormalized Laplacian with the normalized random walk Laplacian (ablation)GLIF on Toxicity · ToxicityTest Acc.: 0.803 ± 0.003 -> 0.803 ± 0.003No clear effect
use GLIF rather than the global IF-constraints post-processingGloVe sentiment model · Sentimentaccuracy at comparable fairness: 89.4% -> 95%Improved
use GLIF post-processingGloVe sentiment model · Sentimentruntime: 422s -> 0.02sImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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