Iterative Semi-Supervised Learning Using Softmax Probability

For the classification problem in practice, one of the challenging issues is to obtain enough labeled data for training. Moreover, even if such labeled data has been sufficiently accumulated, most datasets often exhibit long-tailed distribution with heavy class imbalance, which results in a biased model towards a majority class. To alleviate such class imbalance, semi-supervised learning methods using additional unlabeled data have been considered. However, as a matter of course, the accuracy is much lower than that from supervised learning. In this study, under the assumption that additional unlabeled data is available, we propose the iterative semi-supervised learning algorithms, which iteratively correct the labeling of the extra unlabeled data based on softmax probabilities. The results show that the proposed algorithms provide the accuracy as high as that from the supervised learning. To validate the proposed algorithms, we tested on the two scenarios: with the balanced unlabeled dataset and with the imbalanced unlabeled dataset. Under both scenarios, our proposed semi-supervised learning algorithms provided higher accuracy than previous state-of-the-arts. Code is available at https://github. com/HeewonChung92/iterative- semi-learning.

Automated summary

One of the challenging issues in classification problems is obtaining enough labeled data for training. Most datasets exhibit class imbalance, which biases the model towards the majority class. To address this issue, semi-supervised learning methods using additional unlabeled data have been proposed. In this study, we propose iterative semi-supervised learning algorithms that correct the labeling of extra unlabeled data based on softmax probabilities. The results show that our algorithms achieve high accuracy comparable to supervised learning. Tested on both balanced and imbalanced unlabeled datasets, our algorithms outperform previous state-of-the-art methods.