Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Multi-Objective Interpolation Training for Robustness to Label Noise

About

Deep neural networks trained with standard cross-entropy loss memorize noisy labels, which degrades their performance. Most research to mitigate this memorization proposes new robust classification loss functions. Conversely, we propose a Multi-Objective Interpolation Training (MOIT) approach that jointly exploits contrastive learning and classification to mutually help each other and boost performance against label noise. We show that standard supervised contrastive learning degrades in the presence of label noise and propose an interpolation training strategy to mitigate this behavior. We further propose a novel label noise detection method that exploits the robust feature representations learned via contrastive learning to estimate per-sample soft-labels whose disagreements with the original labels accurately identify noisy samples. This detection allows treating noisy samples as unlabeled and training a classifier in a semi-supervised manner to prevent noise memorization and improve representation learning. We further propose MOIT+, a refinement of MOIT by fine-tuning on detected clean samples. Hyperparameter and ablation studies verify the key components of our method. Experiments on synthetic and real-world noise benchmarks demonstrate that MOIT/MOIT+ achieves state-of-the-art results. Code is available at https://git.io/JI40X.

Diego Ortego, Eric Arazo, Paul Albert, Noel E. O'Connor, Kevin McGuinness• 2020

Related benchmarks

TaskDatasetResultRank
Image ClassificationCIFAR-100 (test)
Accuracy75.9
3518
Image ClassificationMini-Imagenet (test)
Top-1 Accuracy68.28
232
Image ClassificationILSVRC 2012 (val)
Top-1 Accuracy73.8
156
Image ClassificationWebVision mini (val)
Top-1 Accuracy78.76
78
Image ClassificationCIFAR-100 Symmetric Noise (test)
Accuracy75.9
76
Image ClassificationCIFAR-100 (test)
Accuracy (Symmetric 20%)73
72
Image ClassificationCIFAR-10 (test)
Accuracy93.1
68
Image ClassificationCIFAR-10 Symmetric Noise (test)--
64
Image ClassificationCIFAR-100-N--
62
Image ClassificationWebVision (val)
Top-1 Acc78.76
57
Showing 10 of 35 rows

Other info

Code

Follow for update