Log In Sign Up

Coresets for Robust Training of Neural Networks against Noisy Labels

by   Baharan Mirzasoleiman, et al.

Modern neural networks have the capacity to overfit noisy labels frequently found in real-world datasets. Although great progress has been made, existing techniques are limited in providing theoretical guarantees for the performance of the neural networks trained with noisy labels. Here we propose a novel approach with strong theoretical guarantees for robust training of deep networks trained with noisy labels. The key idea behind our method is to select weighted subsets (coresets) of clean data points that provide an approximately low-rank Jacobian matrix. We then prove that gradient descent applied to the subsets do not overfit the noisy labels. Our extensive experiments corroborate our theory and demonstrate that deep networks trained on our subsets achieve a significantly superior performance compared to state-of-the art, e.g., 6 increase in accuracy on CIFAR-10 with 80 accuracy on mini Webvision.


page 1

page 2

page 3

page 4


Investigating Why Contrastive Learning Benefits Robustness Against Label Noise

Self-supervised contrastive learning has recently been shown to be very ...

Understanding and Utilizing Deep Neural Networks Trained with Noisy Labels

Noisy labels are ubiquitous in real-world datasets, which poses a challe...

Adaptive Second Order Coresets for Data-efficient Machine Learning

Training machine learning models on massive datasets incurs substantial ...

Understanding Generalization of Deep Neural Networks Trained with Noisy Labels

Over-parameterized deep neural networks trained by simple first-order me...

Generalization Guarantees for Neural Networks via Harnessing the Low-rank Structure of the Jacobian

Modern neural network architectures often generalize well despite contai...

Co-sampling: Training Robust Networks for Extremely Noisy Supervision

Training robust deep networks is challenging under noisy labels. Current...

Computational and Statistical Tradeoffs in Learning to Rank

For massive and heterogeneous modern datasets, it is of fundamental inte...