AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Controllable Feature Whitening for Hyperparameter-Free Bias Mitigation

Yooshin Cho, Hanbyel Cho, Janghyeon Lee, HyeongGwon Hong, Jaesung Ahn, Junmo Kim

ICCV 2025 · Label-efficient and robust learning › Recognition and long-tail · approach: Data augmentation and generalization bias · 0 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
add CFW (lambda=0.25) (ablation)Vanilla · bFFHQunbiased accuracy: 77.96 -> 83.9Improved
add CFW (lambda=0.25) (ablation)Vanilla · bFFHQbias-conflicting accuracy: 56.87 -> 78.1Improved
add loss weighting to Vanilla network (ablation)Vanilla · bFFHQunbiased accuracy: 77.96 -> 78.5No clear effect
add loss weighting on top of CFW lambda=0.25 (ablation)CFW · bFFHQunbiased accuracy: 83.9 -> 84.6Improved
use unbiased covariance weighting lambda=0.25 instead of 0.0 (ablation)CFW · bFFHQunbiased accuracy: 80.0 -> 83.9Improved
use lambda=0.25 instead of 0.0 (ablation)CFW · bFFHQbias-conflicting accuracy: 80.6 -> 78.1Hurt
use coupled Newton-Schulz T=5 instead of ZCA (ablation)CFW · bFFHQbias-conflicting accuracy: 72.66 -> 79.8Improved
use cNS T=5 instead of Cholesky decomposition (ablation)CFW · bFFHQunbiased accuracy: 74.06 -> 84.6Improved
use cNS T=5 instead of T=3 (ablation)CFW · bFFHQbias-conflicting accuracy: 76.80 -> 79.8Improved
use SelecMix pretrained target encoder instead of Vanilla (ablation)Ours · bFFHQbias-conflicting accuracy: 79.8 -> 82.77Improved
use SelecMix pretrained target encoder instead of Vanilla (ablation)Ours · Corrupted CIFAR-10 0.5%unbiased accuracy: 32.08 -> 42.51Improved
use CFW instead of adversarial GRLOurs · Celeb-A T=attractive/S=maleequalized odds violation: 24.9 -> 4.2Improved

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