All Articles

Towards Galaxy Foundation Models

This blog is a summary of my ICML Machine Learning for Astrophysics paper “Towards Galaxy Foundation Models with Hybrid Contrastive Learning”.

Computer science is coming to terms with a result that’s both exciting and tragic. For nearly any task, training a model directly on that task works less well than finetuning some existing large model that was pretrained on a huge and diverse dataset. These models are (controversially) called foundation models.

This is exciting because we can build better models than before - models which are not only more accurate but also more robust. It’s tragic because it risks entrenching well-resourced labs as the highly-paid gatekeepers of those models. But first let’s consider what this means for you.

My impression of our field is that many extragalactic astronomers:

  • Have a specific galaxy morphology question not answered by GZ’s questions (e.g. finding a rare galaxy type)
  • Have 100-10,000 expert-labelled galaxies where the answer is known
  • Want to solve the problem with deep learning (for effectiveness and for hype/career impact) but are not DL experts

Foundation models are perfect for this. Thanks to pretraining, they can be adapted to new tasks using very little data. I want to create foundation models for galaxies using Galaxy Zoo.

Finetuning pretrained "foundation" models can answer new questions with minimal data. These ring galaxies were found by finetuning Zoobot, a CNN pretrained on those questions, with ~200 examples of rings.

I previously open-sourced Zoobot - a CNN pretrained to solve every GZ DECaLS DR5 question. My tests showed that adapting Zoobot worked better than adapting from ImageNet or training the same model directly.

This new paper takes the same idea and scales it up. I make two key changes:

  1. More labels: I train on four separate Galaxy Zoo campaigns, with nearly 100M human clicks between them.
  2. More images: I use contrastive learning to benefit from 1.3M images that haven’t been labelled by humans.

Training on different Galaxy Zoo campaigns is tricky because the questions and answers are different between campaigns and images are usually only labelled in a single campaign. So what should the model predict? Happily, the Dirichlet loss function I introduced in GZ DECaLS can handle this naturally. The loss essentially measures p(kw,N)p(k|w, N) i.e. the odds of kk out of NN volunteers giving some answer to some question and for some deep learning model ww. But when NN is 0, p(k=0N=0,w)=1p(k=0|N=0, w) = 1 and so the w doesn’t affect the loss - hence unanswered questions have no effect on training. It’s therefore easy to learn from galaxies with only a few answered questions: if the question is not answered, the model simply ignores it.

To train on totally unlabelled images - where no questions in any GZ campaign have been answered - I used the contrastive learning framework Bootstrap Your Own Latent (BYOL). This essentially presents two randomly-augmented versions of the same image to two networks, and asks one network to predict the internal representation of the other network. Because they see differently-augmented images, the representations of both must be invariant to those augmentations, which provides a learning signal. But BYOL is unsupervised, and I want to benefit from both unlabelled and labelled images. I did this by adding a supervised prediction head to one of BYOL’s networks, forcing it to both predict the representation of another network and solve the supervised Dirichlet loss task at the same time.

Adding a new supervised head (top) to BYOL (Grill 2020) helps guide the otherwise-unsupervised method to solve downstream tasks astronomers care about.

Making these two changes creates a model which substantially outperforms both direct training and pretraining, for the specific task of finding ring galaxies given scarce labels.

Historically, Galaxy Zoo has shared catalogs of vote counts for questions we hoped would be generally useful to many people. Choosing these questions is a compromise between scientific precision, broad applicability, and limited volunteer time. Now we no longer need to compromise. I believe Galaxy Zoo can shift towards creating adaptable foundation models that you can use to answer exactly the questions you care about.

I’d love to test this on more than just ring galaxies. If you have a small expert-labelled dataset centered on a specific question, please reach out (email, twitter).