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
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 changed | Where | Measured | Effect |
|---|---|---|---|
| search architecture and hyperparameters with NAS+HPO instead of standard architecture | face recognition (ArcFace baseline) · CelebA | error (%): 4.35 -> 3.10 | Improved |
| search architecture with NAS+HPO instead of SOTA baseline | face recognition · VGGFace2 | error (%): 4.5 -> 3.66 | Improved |
| NAS+HPO model instead of DPN | DPN · VGGFace2 | rank disparity: 0.27 -> 0.23 | Improved |
| apply Flipped mitigation on top of SMAC_000 (ablation) | SMAC_000 · CelebA | rank disparity: 2.18 -> 0.03 | Improved |
| apply Flipped on top of SMAC_680 (ablation) | SMAC_680 · CelebA | rank disparity: 1.96 -> 4.50 | Hurt |
| apply Flipped on top of SMAC_301 (ablation) | SMAC_301 · VGGFace2 | rank disparity: 0.23 -> 0.18 | Improved |
| apply SensitiveNets on top of SMAC_301 (ablation) | SMAC_301 · VGGFace2 | rank disparity: 0.23 -> 0.41 | Hurt |
| apply Angular margin on top of SMAC_301 (ablation) | SMAC_301 · VGGFace2 | rank disparity: 0.23 -> 0.25 | Hurt |
| use SMAC instead of MO-ASHA for search (ablation) | NAS+HPO · VGGFace2 | rank disparity: 0.408 -> 0.230 | Improved |
| use SMAC instead of NSGA-II for search (ablation) | NAS+HPO · VGGFace2 | accuracy: 86.811 -> 96.337 | Improved |
| transfer SMAC_301 vs DPN_SGD to other datasets | SMAC_301 · LFW | accuracy: 93.0 -> 96.63 | Improved |
| use searched architecture instead of DPN | SMAC_301 · VGGFace2 linear probe | gender accuracy (%): 90.719 -> 69.980 | Improved |
Closest papers by meaning
- Mitigating Gender Bias in Face Recognition using the von Mises-Fisher Mixture Model (ICML 2022 · similarity 0.91)
- Towards Fairness in Visual Recognition: Effective Strategies for Bias Mitigation (CVPR 2020 · similarity 0.90)
- Benchmarking Algorithmic Bias in Face Recognition: An Experimental Approach Using Synthetic Faces and Human Evaluation (ICCV 2023 · similarity 0.89)
- Classes Are Not Equal: An Empirical Study on Image Recognition Fairness (CVPR 2024 · similarity 0.87)
- FAIR-FER: A Latent Alignment Approach for Mitigating Bias in Facial Expression Recognition (Student Abstract) (AAAI 2024 · similarity 0.86)
- Leveraging Diffusion Perturbations for Measuring Fairness in Computer Vision (AAAI 2024 · similarity 0.86)
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.
