Papers › PSSCL: A progressive sample selection framework with contrastive loss designed for noisy labels

PSSCL: A progressive sample selection framework with contrastive loss designed for noisy labels

18 Dec 2024Pattern Recognition 2024 12archive 2025-07-28

Qian Zhang, Yi Zhu, Filipe R. Cordeiro, Qiu Chen

Large-scale image datasets frequently contain unavoidable noisy labels, resulting in overfitting in deep neural networks and declining performance. Most existing methods for learning from noisy labels operate as one-stage frameworks, where training data division and semi-supervised learning (SSL) are intertwined for optimization. Accordingly, their effectiveness is significantly influenced by the precision of the separated clean set, prior knowledge of noise, and the robustness of SSL. In this paper, we propose a progressive sample selection framework with contrastive loss for noisy labels named PSSCL. This framework operates in two stages, using robust and contrastive losses to augment the robustness of the model. Stage I focuses on identifying a small clean set through a long-term confidence detection strategy, while stage II aims to enhance performance by expanding this clean set. PSSCL demonstrates significant improvement across various benchmarks when compared with state-of-the-art methods. The code is available at https://github.com/LanXiaoPang613/PSSCL.

PaperPDFCode

Code

LanXiaoPang613/PSSCL mentioned in paperpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Image ClassificationLearning with noisy labels

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification mini WebVision 1.0 PSSCL (130 epochs) ImageNet Top-1 Accuracy 79.68 #13 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PSSCL (130 epochs) ImageNet Top-5 Accuracy 95.16 #13 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PSSCL (130 epochs) Top-1 Accuracy 79.56 #13 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PSSCL (130 epochs) Top-5 Accuracy 94.84 #13 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PSSCL (120 epochs) ImageNet Top-1 Accuracy 79.40 #23 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PSSCL (120 epochs) ImageNet Top-5 Accuracy 94.84 #23 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PSSCL (120 epochs) Top-1 Accuracy 78.52 #23 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PSSCL (120 epochs) Top-5 Accuracy 93.80 #23 of 47 Archive leaderboard report
Learning with noisy labels ANIMAL PSSCL Accuracy 88.74 #4 of 19 Archive leaderboard report
Learning with noisy labels ANIMAL PSSCL ImageNet Pretrained NO #4 of 19 Archive leaderboard report
Learning with noisy labels ANIMAL PSSCL Network Vgg19-BN #4 of 19 Archive leaderboard report
Learning with noisy labels CIFAR-100N PSSCL Accuracy (mean) 72.00 #3 of 24 Archive leaderboard report
Learning with noisy labels CIFAR-10N-Aggregate PSSCL Accuracy (mean) 96.41 #2 of 26 Archive leaderboard report
Learning with noisy labels CIFAR-10N-Random1 PSSCL Accuracy (mean) 96.17 #2 of 24 Archive leaderboard report
Learning with noisy labels CIFAR-10N-Random2 PSSCL Accuracy (mean) 96.21 #1 of 23 Archive leaderboard report
Learning with noisy labels CIFAR-10N-Random3 PSSCL Accuracy (mean) 96.49 #1 of 23 Archive leaderboard report
Learning with noisy labels CIFAR-10N-Worst PSSCL Accuracy (mean) 95.12 #2 of 25 Archive leaderboard report
Learning with noisy labels Food-101 PSSCL Accuracy (% ) 86.41 #2 of 3 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

SET

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections