# Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learning

## TL;DR

This paper explores pseudo-labeling in semi-supervised image classification, identifying confirmation bias as a challenge and proposing simple regularization techniques that outperform more complex consistency regularization methods.

## Contribution

It demonstrates that pseudo-labeling, with regularization like mixup and minimum labeled samples, can surpass consistency regularization in semi-supervised learning.

## Key findings

- Pseudo-labeling can outperform consistency regularization methods.
- Mixup augmentation reduces confirmation bias.
- Achieved state-of-the-art results on multiple datasets.

## Abstract

Semi-supervised learning, i.e. jointly learning from labeled and unlabeled samples, is an active research topic due to its key role on relaxing human supervision. In the context of image classification, recent advances to learn from unlabeled samples are mainly focused on consistency regularization methods that encourage invariant predictions for different perturbations of unlabeled samples. We, conversely, propose to learn from unlabeled data by generating soft pseudo-labels using the network predictions. We show that a naive pseudo-labeling overfits to incorrect pseudo-labels due to the so-called confirmation bias and demonstrate that mixup augmentation and setting a minimum number of labeled samples per mini-batch are effective regularization techniques for reducing it. The proposed approach achieves state-of-the-art results in CIFAR-10/100, SVHN, and Mini-ImageNet despite being much simpler than other methods. These results demonstrate that pseudo-labeling alone can outperform consistency regularization methods, while the opposite was supposed in previous work. Source code is available at https://git.io/fjQsC.