Research map / Learning theory and algorithms / Fairness
FairGRAPE: Fairness-Aware GRAdient Pruning mEthod for Face Attribute Classification
Xiaofeng Lin, Seungbae Kim, Jungseock Joo
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 |
|---|---|---|---|
| add group-wise importance and fair node selection to pruning | FairGRAPE · FairFace race (99% sparsity) | accuracy: 62.3 -> 66.8 | Improved |
| add group importance (ablation) | FairGRAPE · CelebA attributes | accuracy: 89.31 -> 90.90 | Improved |
| add group importance (ablation) | FairGRAPE · CelebA attributes | std of accuracy loss: 0.31 -> 0.11 | Improved |
| add iterative retraining (ablation) | FairGRAPE · CelebA attributes | accuracy: 89.26 -> 90.90 | Improved |
| increase pruning step r from 0.1 to 0.2 (ablation) | FairGRAPE · CelebA attributes | accuracy: 90.90 -> 90.49 | Hurt |
| increase pruning step r to 0.5 (ablation) | FairGRAPE · CelebA attributes | accuracy: 90.90 -> 90.34 | Hurt |
| use 100% instead of 20% training images for importance (ablation) | FairGRAPE · CelebA attributes | accuracy: 90.90 -> 90.66 | Hurt |
| compute pruning on minority-race examples only (ablation) | GraSP · UTKFace gender | std of accuracy loss: 1.77 -> 2.31 | Hurt |
| compute pruning on minority-race examples only (ablation) | SNIP · UTKFace gender | std of accuracy loss: 1.62 -> 1.90 | Hurt |
| compute pruning on minority-race examples only (ablation) | SNIP · UTKFace gender | overall accuracy: 90.95 -> 90.55 | No clear effect |
| use K-means clusters instead of race labels as groups | FairGRAPE · FairFace race | std of accuracy loss: 11.3 -> 7.40 | Improved |
| apply FairGRAPE instead of GraSP on ImageNet person subtree | FairGRAPE · ImageNet person subtree | accuracy: 15.36 -> 51.12 | Improved |
Closest papers by meaning
- On Demographic Group Fairness Guarantees in Deep Learning (TPAMI 2026 · similarity 0.86)
- Mitigating Demographic Bias in Facial Datasets with Style-Based Multi-attribute Transfer (IJCV 2021 · similarity 0.86)
- FairSIN: Achieving Fairness in Graph Neural Networks through Sensitive Information Neutralization (AAAI 2024 · similarity 0.86)
- FairNet: Dynamic Fairness Correction without Performance Loss via Contrastive Conditional LoRA (NeurIPS 2025 · similarity 0.85)
- GDeR: Safeguarding Efficiency, Balancing, and Robustness via Prototypical Graph Pruning (NeurIPS 2024 · similarity 0.84)
- Distributionally Generative Augmentation for Fair Facial Attribute Classification (CVPR 2024 · similarity 0.84)
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.
