An automatic taxonomy of galaxy morphology using unsupervised machine learning

TL;DR

This paper introduces an unsupervised machine learning method that automatically classifies galaxy morphologies from imaging data without pre-selection, demonstrating its effectiveness across multiple surveys and matching human classifications.

Contribution

The authors develop a novel unsupervised technique for galaxy classification that does not require pre-labeled data or pre-filtering, enabling scalable and unbiased morphological analysis.

Findings

Successfully classified galaxies in multiple HST fields

Achieved good concordance with Galaxy Zoo classifications

Identified rare objects and new lensed galaxy candidates

Abstract

We present an unsupervised machine learning technique that automatically segments and labels galaxies in astronomical imaging surveys using only pixel data. Distinct from previous unsupervised machine learning approaches used in astronomy we use no pre-selection or pre-filtering of target galaxy type to identify galaxies that are similar. We demonstrate the technique on the HST Frontier Fields. By training the algorithm using galaxies from one field (Abell 2744) and applying the result to another (MACS0416.1-2403), we show how the algorithm can cleanly separate early and late type galaxies without any form of pre-directed training for what an 'early' or 'late' type galaxy is. We then apply the technique to the HST CANDELS fields, creating a catalogue of approximately 60,000 classifications. We show how the automatic classification groups galaxies of similar morphological (and photometric) type, and make the classifications public via a catalogue, a visual catalogue and galaxy similarity search. We compare the CANDELS machine-based classifications to human-based classifications from the Galaxy Zoo: CANDELS project. Although there is not a direct mapping between Galaxy Zoo and our hierarchical labelling, we demonstrate a good level of concordance between human and machine classifications. Finally, we show how the technique can be used to identify rarer objects and present new lensed galaxy candidates from the CANDELS imaging.