AI Paper Atlas

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

Batch Loss Score for Dynamic Data Pruning

Qing Zhou, Bingxuan Zhao, Tao Yang, Hongyuan Zhang, Junyu Gao, Qi Wang

CVPR 2026 · Label-efficient and robust learning › Noisy labels · approach: Contrastive learning on noisy data · 2 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 EMA smoothing to batch-loss score (ablation)BLS-InfoBatch · CIFAR100 ResNet18accuracy (%): 77.9 -> 78.3Improved
add EMA smoothing to batch-loss score (ablation)BLS-SeTa · CIFAR100 ResNet50accuracy (%): 79.9 -> 80.8Improved
add EMA smoothing to batch-loss score (ablation)BLS-InfoBatch · CIFAR100 ResNet50accuracy (%): 79.9 -> 80.6Improved
replace per-sample loss with BLS in InfoBatchInfoBatch · CIFAR100 ResNet18 30% pruningaccuracy (%): 78.2 -> 78.4No clear effect
replace per-sample loss with BLS in SeTaSeTa · NoCaps Val Overall (ToCa)CIDEr: 71.5 -> 71.2No clear effect
replace per-sample loss with BLS in InfoBatchInfoBatch · NoCaps Val Overall (ToCa)CIDEr: 70.2 -> 70.6No clear effect
replace per-sample loss with BLS in InfoBatch at 50% pruningInfoBatch · CIFAR100 ResNet18 50% pruningaccuracy (%): 78.1 -> 77.7No clear effect
prune 22.9% of ImageNet-1K with BLS-InfoBatchBLS-InfoBatch · ImageNet-1K Vimaccuracy (%): 75.7 -> 75.6No clear effect
BLS-SeTa pruning on YOLOv5n segmentationBLS-SeTa · COCO instance segmentationmAP.5: 29.5 -> 31.7Improved
BLS-InfoBatch pruning on CIFAR100 classification with YOLOv5nBLS-InfoBatch · CIFAR100 YOLOv5nAcc@1: 58.0 -> 57.6Hurt
raise EMA decay alpha from 0.5 to 0.7-0.8 (ablation)BLS · CIFAR100 ResNet18accuracy (%): 79.0-79.1 -> 79.4Improved
set EMA decay alpha to 1.0 (ablation)BLS · CIFAR100 ResNet18/50pruning ratioHurt

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