A Contrastive Learning Approach to Auroral Identification and Classification

TL;DR

This paper applies a contrastive learning method to auroral image classification, achieving high accuracy with fewer parameters and revealing more detailed clustering than existing categories, thus enhancing practical auroral analysis.

Contribution

It adapts the SimCLR contrastive learning framework to auroral images, significantly improving classification accuracy and uncovering finer distinctions in auroral types compared to traditional methods.

Findings

Achieved nearly 10% higher classification accuracy than previous benchmarks.

Learned representations naturally form more clusters than manual categories.

Model requires fewer than 25% of the parameters of previous benchmarks.

Abstract

Unsupervised learning algorithms are beginning to achieve accuracies comparable to their supervised counterparts on benchmark computer vision tasks, but their utility for practical applications has not yet been demonstrated. In this work, we present a novel application of unsupervised learning to the task of auroral image classification. Specifically, we modify and adapt the Simple framework for Contrastive Learning of Representations (SimCLR) algorithm to learn representations of auroral images in a recently released auroral image dataset constructed using image data from Time History of Events and Macroscale Interactions during Substorms (THEMIS) all-sky imagers. We demonstrate that (a) simple linear classifiers fit to the learned representations of the images achieve state-of-the-art classification performance, improving the classification accuracy by almost 10 percentage points over the current benchmark; and (b) the learned representations naturally cluster into more clusters than exist manually assigned categories, suggesting that existing categorizations are overly coarse and may obscure important connections between auroral types, near-earth solar wind conditions, and geomagnetic disturbances at the earth's surface. Moreover, our model is much lighter than the previous benchmark on this dataset, requiring in the area of fewer than 25\% of the number of parameters. Our approach exceeds an established threshold for operational purposes, demonstrating readiness for deployment and utilization.