AI Paper Atlas

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

Combating Label Noise with a General Surrogate Model for Sample Selection

Chao Liang, Linchao Zhu, Humphrey Shi, Yi Yang

IJCV 2024 · Label-efficient and robust learning › Noisy labels · approach: Sample selection and co-training · 8 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 CLIP-based clean sample selection and new lossDivideMix · WebVisiontop-1 accuracy: 77.32 -> 79.08Improved
add CLIP-based selection and lossDivideMix · CIFAR-10 sym 90%accuracy: 75.4 -> 89.2Improved
add CLIP-based selection and lossDivideMix · CIFAR-100 sym 90%accuracy: 31.0 -> 45.7Improved
add CLIP-based selection and lossDivideMix · CIFAR-10 asym 40%accuracy: 92.1 -> 95.1Improved
add CLIP-based selection and lossDivideMix · Clothing-1Maccuracy: 74.76 -> 74.84No clear effect
add CLIP-based selectionInstanceGM · Red Mini-ImageNet 80%accuracy: 39.62 -> 45.65Improved
select samples by CLIP prediction confidence instead of GMM small loss (ablation)sampling strategy · WebVisiontop-1 accuracy: 77.36 -> 78.32Improved
select by CLIP prompt consistency instead of prediction confidence (ablation)sampling strategy · WebVisiontop-1 accuracy: 78.32 -> 77.36Hurt
add focal loss instead of cross-entropy (ablation)classifier training · WebVisiontop-1 accuracy: 77.76 -> 78.32Improved
add focal loss to full loss (ablation)noise-aware balanced margin loss · WebVisiontop-1 accuracy: 78.20 -> 79.08Improved
add balanced margin (ablation)noise-aware balanced margin loss · WebVisiontop-1 accuracy: 78.68 -> 79.08Improved
add noise-aware margin (ablation)noise-aware balanced margin loss · WebVisiontop-1 accuracy: 78.76 -> 79.08Improved

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