Practical Galaxy Morphology Tools from Deep Supervised Representation Learning

TL;DR

Deep supervised models trained on Galaxy Zoo data learn meaningful galaxy representations that enable accurate similarity searches, anomaly detection, and efficient adaptation to new tasks with minimal additional labels, challenging the need for large new datasets.

Contribution

The paper demonstrates that deep supervised models trained on existing galaxy data produce transferable representations useful for various practical tasks in astronomy.

Findings

Models outperform recent approaches in galaxy similarity search.

100% accuracy in identifying top anomalies judged by volunteers.

Fine-tuned models outperform ImageNet-based models and training from scratch in new tasks.

Abstract

Astronomers have typically set out to solve supervised machine learning problems by creating their own representations from scratch. We show that deep learning models trained to answer every Galaxy Zoo DECaLS question learn meaningful semantic representations of galaxies that are useful for new tasks on which the models were never trained. We exploit these representations to outperform several recent approaches at practical tasks crucial for investigating large galaxy samples. The first task is identifying galaxies of similar morphology to a query galaxy. Given a single galaxy assigned a free text tag by humans (e.g. "#diffuse"), we can find galaxies matching that tag for most tags. The second task is identifying the most interesting anomalies to a particular researcher. Our approach is 100% accurate at identifying the most interesting 100 anomalies (as judged by Galaxy Zoo 2 volunteers). The third task is adapting a model to solve a new task using only a small number of newly-labelled galaxies. Models fine-tuned from our representation are better able to identify ring galaxies than models fine-tuned from terrestrial images (ImageNet) or trained from scratch. We solve each task with very few new labels; either one (for the similarity search) or several hundred (for anomaly detection or fine-tuning). This challenges the longstanding view that deep supervised methods require new large labelled datasets for practical use in astronomy. To help the community benefit from our pretrained models, we release our fine-tuning code Zoobot. Zoobot is accessible to researchers with no prior experience in deep learning.