AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Rethinking Bias Mitigation: Fairer Architectures Make for Fairer Face Recognition

Samuel Dooley, Rhea Sanjay Sukthanker, John P Dickerson, Colin White, Frank Hutter, Micah Goldblum

NeurIPS 2023 · Architectures and efficiency › Neural architecture search · approach: AutoML pipeline search · 9 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
search architecture and hyperparameters with NAS+HPO instead of standard architectureface recognition (ArcFace baseline) · CelebAerror (%): 4.35 -> 3.10Improved
search architecture with NAS+HPO instead of SOTA baselineface recognition · VGGFace2error (%): 4.5 -> 3.66Improved
NAS+HPO model instead of DPNDPN · VGGFace2rank disparity: 0.27 -> 0.23Improved
apply Flipped mitigation on top of SMAC_000 (ablation)SMAC_000 · CelebArank disparity: 2.18 -> 0.03Improved
apply Flipped on top of SMAC_680 (ablation)SMAC_680 · CelebArank disparity: 1.96 -> 4.50Hurt
apply Flipped on top of SMAC_301 (ablation)SMAC_301 · VGGFace2rank disparity: 0.23 -> 0.18Improved
apply SensitiveNets on top of SMAC_301 (ablation)SMAC_301 · VGGFace2rank disparity: 0.23 -> 0.41Hurt
apply Angular margin on top of SMAC_301 (ablation)SMAC_301 · VGGFace2rank disparity: 0.23 -> 0.25Hurt
use SMAC instead of MO-ASHA for search (ablation)NAS+HPO · VGGFace2rank disparity: 0.408 -> 0.230Improved
use SMAC instead of NSGA-II for search (ablation)NAS+HPO · VGGFace2accuracy: 86.811 -> 96.337Improved
transfer SMAC_301 vs DPN_SGD to other datasetsSMAC_301 · LFWaccuracy: 93.0 -> 96.63Improved
use searched architecture instead of DPNSMAC_301 · VGGFace2 linear probegender accuracy (%): 90.719 -> 69.980Improved

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