Meta Soft Label Generation for Noisy Labels

TL;DR

This paper introduces MSLG, a meta-learning based method for generating soft labels to mitigate the impact of noisy labels in training deep neural networks, improving performance across multiple datasets.

Contribution

We propose a novel meta soft label generation algorithm that adaptively estimates optimal labels using meta-learning, applicable to any model and effective against noisy labels.

Findings

MSLG outperforms state-of-the-art methods on CIFAR10, Clothing1M, and Food101N.

The approach effectively reduces the impact of noisy labels in deep learning.

MSLG is model-agnostic and can be integrated with existing models easily.

Abstract

The existence of noisy labels in the dataset causes significant performance degradation for deep neural networks (DNNs). To address this problem, we propose a Meta Soft Label Generation algorithm called MSLG, which can jointly generate soft labels using meta-learning techniques and learn DNN parameters in an end-to-end fashion. Our approach adapts the meta-learning paradigm to estimate optimal label distribution by checking gradient directions on both noisy training data and noise-free meta-data. In order to iteratively update soft labels, meta-gradient descent step is performed on estimated labels, which would minimize the loss of noise-free meta samples. In each iteration, the base classifier is trained on estimated meta labels. MSLG is model-agnostic and can be added on top of any existing model at hand with ease. We performed extensive experiments on CIFAR10, Clothing1M and Food101N datasets. Results show that our approach outperforms other state-of-the-art methods by a large margin.