Papers › Sample Prior Guided Robust Model Learning to Suppress Noisy Labels

Sample Prior Guided Robust Model Learning to Suppress Noisy Labels

2 Dec 2021arXiv:2112.01197archive 2025-07-28

Wenkai Chen, Chuang Zhu, Yi Chen, Mengting Li, Tiejun Huang

Imperfect labels are ubiquitous in real-world datasets and seriously harm the model performance. Several recent effective methods for handling noisy labels have two key steps: 1) dividing samples into cleanly labeled and wrongly labeled sets by training loss, 2) using semi-supervised methods to generate pseudo-labels for samples in the wrongly labeled set. However, current methods always hurt the informative hard samples due to the similar loss distribution between the hard samples and the noisy ones. In this paper, we proposed PGDF (Prior Guided Denoising Framework), a novel framework to learn a deep model to suppress noise by generating the samples' prior knowledge, which is integrated into both dividing samples step and semi-supervised step. Our framework can save more informative hard clean samples into the cleanly labeled set. Besides, our framework also promotes the quality of pseudo-labels during the semi-supervised step by suppressing the noise in the current pseudo-labels generating scheme. To further enhance the hard samples, we reweight the samples in the cleanly labeled set during training. We evaluated our method using synthetic datasets based on CIFAR-10 and CIFAR-100, as well as on the real-world datasets WebVision and Clothing1M. The results demonstrate substantial improvements over state-of-the-art methods.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

bupt-ai-cz/PGDF officialmentioned on GitHubpytorch 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 CIFAR-10 (with noisy labels) PGDF (ResNet-18) Accuracy (under 20% Sym. label noise) 96.7% #3 of 4 Archive leaderboard report
Image Classification CIFAR-10 (with noisy labels) PGDF (ResNet-18) Accuracy (under 50% Sym. label noise) 96.3% #3 of 4 Archive leaderboard report
Image Classification CIFAR-10 (with noisy labels) PGDF (ResNet-18) Accuracy (under 80% Sym. label noise) 94.7% #3 of 4 Archive leaderboard report
Image Classification CIFAR-10 (with noisy labels) PGDF (ResNet-18) Accuracy (under 90% Sym. label noise) 84.0% #3 of 4 Archive leaderboard report
Image Classification Clothing1M PGDF Accuracy 75.19% #9 of 51 Archive leaderboard report
Image Classification mini WebVision 1.0 PGDF (Inception-ResNet-v2) ImageNet Top-1 Accuracy 75.45 #3 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PGDF (Inception-ResNet-v2) ImageNet Top-5 Accuracy 93.11 #3 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PGDF (Inception-ResNet-v2) Top-1 Accuracy 81.47 #3 of 47 Archive leaderboard report
Image Classification mini WebVision 1.0 PGDF (Inception-ResNet-v2) Top-5 Accuracy 94.03 #3 of 47 Archive leaderboard report
Learning with noisy labels CIFAR-100N PGDF Accuracy (mean) 74.08 #1 of 24 Archive leaderboard report
Learning with noisy labels CIFAR-10N-Aggregate PGDF Accuracy (mean) 96.11 #3 of 26 Archive leaderboard report
Learning with noisy labels CIFAR-10N-Random1 PGDF Accuracy (mean) 96.01 #3 of 24 Archive leaderboard report
Learning with noisy labels CIFAR-10N-Worst PGDF Accuracy (mean) 93.65 #3 of 25 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.

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