HIERMATCH: Leveraging Label Hierarchies for Improving Semi-Supervised Learning

TL;DR

HIERMATCH is a novel semi-supervised learning framework that leverages hierarchical label information to reduce labeling costs and improve performance, applicable to existing methods like MixMatch and FixMatch.

Contribution

It introduces a generic approach that utilizes coarse labels as weak supervision to enhance semi-supervised learning, reducing the need for fine-grained labels.

Findings

Reduces fine-grained label usage by 50% on CIFAR-100

Achieves only 0.59% accuracy drop compared to state-of-the-art methods

Demonstrates effectiveness on CIFAR-100 and NABirds datasets

Abstract

Semi-supervised learning approaches have emerged as an active area of research to combat the challenge of obtaining large amounts of annotated data. Towards the goal of improving the performance of semi-supervised learning methods, we propose a novel framework, HIERMATCH, a semi-supervised approach that leverages hierarchical information to reduce labeling costs and performs as well as a vanilla semi-supervised learning method. Hierarchical information is often available as prior knowledge in the form of coarse labels (e.g., woodpeckers) for images with fine-grained labels (e.g., downy woodpeckers or golden-fronted woodpeckers). However, the use of supervision using coarse category labels to improve semi-supervised techniques has not been explored. In the absence of fine-grained labels, HIERMATCH exploits the label hierarchy and uses coarse class labels as a weak supervisory signal. Additionally, HIERMATCH is a generic-approach to improve any semisupervised learning framework, we demonstrate this using our results on recent state-of-the-art techniques MixMatch and FixMatch. We evaluate the efficacy of HIERMATCH on two benchmark datasets, namely CIFAR-100 and NABirds. HIERMATCH can reduce the usage of fine-grained labels by 50% on CIFAR-100 with only a marginal drop of 0.59% in top-1 accuracy as compared to MixMatch. Code: https://github.com/07Agarg/HIERMATCH