Empirical Perspectives on One-Shot Semi-supervised Learning

TL;DR

This paper empirically examines one-shot semi-supervised learning, focusing on factors affecting accuracy and reliability, and explores potential solutions to improve deep neural network training with minimal labeled data.

Contribution

It provides an empirical analysis of the challenges in one-shot semi-supervised learning and investigates factors impacting performance, especially for CIFAR-10.

Findings

Uneven class accuracy hinders high-performance learning

Factors influencing accuracy include data distribution and training dynamics

Potential solutions could enable broader adoption of one-shot semi-supervised methods

Abstract

One of the greatest obstacles in the adoption of deep neural networks for new applications is that training the network typically requires a large number of manually labeled training samples. We empirically investigate the scenario where one has access to large amounts of unlabeled data but require labeling only a single prototypical sample per class in order to train a deep network (i.e., one-shot semi-supervised learning). Specifically, we investigate the recent results reported in FixMatch for one-shot semi-supervised learning to understand the factors that affect and impede high accuracies and reliability for one-shot semi-supervised learning of Cifar-10. For example, we discover that one barrier to one-shot semi-supervised learning for high-performance image classification is the unevenness of class accuracy during the training. These results point to solutions that might enable more widespread adoption of one-shot semi-supervised training methods for new applications.