PENCIL: Deep Learning with Noisy Labels

TL;DR

PENCIL is a versatile deep learning framework that effectively handles noisy labels by jointly updating network parameters and label distributions, improving robustness and accuracy across various datasets and tasks.

Contribution

It introduces a noise-robust, end-to-end method that does not require clean data or prior noise information, applicable to multiple network structures and tasks.

Findings

Outperforms state-of-the-art methods on synthetic and real-world noisy datasets.

Effective in multi-label classification with added attention mechanisms.

Maintains robustness even on clean datasets.

Abstract

Deep learning has achieved excellent performance in various computer vision tasks, but requires a lot of training examples with clean labels. It is easy to collect a dataset with noisy labels, but such noise makes networks overfit seriously and accuracies drop dramatically. To address this problem, we propose an end-to-end framework called PENCIL, which can update both network parameters and label estimations as label distributions. PENCIL is independent of the backbone network structure and does not need an auxiliary clean dataset or prior information about noise, thus it is more general and robust than existing methods and is easy to apply. PENCIL can even be used repeatedly to obtain better performance. PENCIL outperforms previous state-of-the-art methods by large margins on both synthetic and real-world datasets with different noise types and noise rates. And PENCIL is also effective in multi-label classification tasks through adding a simple attention structure on backbone networks. Experiments show that PENCIL is robust on clean datasets, too.