AI Paper Atlas

Research map / Label-efficient and robust learning / Noisy labels

BadLabel: A Robust Perspective on Evaluating and Enhancing Label-Noise Learning

Jingfeng Zhang, Bo Song, Haohan Wang, Bo Han, Tongliang Liu, Lei Liu, Masashi Sugiyama

TPAMI 2024 · Label-efficient and robust learning › Noisy labels · approach: Noise-aware loss and pairwise methods · 45 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
replace DivideMix with Robust DivideMixDivideMix · CIFAR-10 BadLabel 60%best test accuracy: 28.38 -> 76.47Improved
replace DivideMix with Robust DivideMixDivideMix · CIFAR-100 BadLabel 80%best test accuracy: 4.67 -> 21.48Improved
replace DivideMix with Robust DivideMixDivideMix · CIFAR-10 symmetric 80%best test accuracy: 81.95 -> 61.59Hurt
replace DivideMix with Robust DivideMixDivideMix · Clothing1Mtest accuracy: 74.76 -> 74.13No clear effect
replace DivideMix with Robust DivideMixDivideMix · CIFAR-10N (Worst)test accuracy: 92.56 -> 92.70No clear effect
add BayesGMM instead of classical GMM (ablation)Robust DivideMix · CIFAR-10 BadLabel 60%best test accuracy: 44.35 -> 76.47Improved
add label perturbation (ablation)Robust DivideMix · CIFAR-10 BadLabel 40%best test accuracy: 60.16 -> 86.70Improved
add label perturbation (ablation)Robust DivideMix · CIFAR-10 BadLabel 60%best test accuracy: 67.98 -> 76.47Improved
add filtering of low-quality divisions (ablation)Robust DivideMix · CIFAR-10 BadLabel 20%best test accuracy: 75.19 -> 92.07Improved
add filtering of low-quality divisions (ablation)Robust DivideMix · CIFAR-100 BadLabel 80%best test accuracy: 5.38 -> 21.48Improved
add BayesGMM instead of classical GMM (ablation)Robust DivideMix · CIFAR-100 BadLabel 60%best test accuracy: 30.18 -> 41.80Improved
raise label perturbation step lambda from 0.2 to 1.5 on CIFAR-10 symmetric noise (ablation)Robust DivideMix · CIFAR-10 symmetricbest test accuracy: 86.53 -> 64.22Hurt

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