Intro to Foundation Models for Astrophysics

Francois Lanusse

Paris Workshop on Bayesian Deep Learning, May 2025

https://slides.com/eiffl/fm-tuto

  • 1000 images each night, 15 TB/night for 10 years

  • 18,000 square degrees, observed once every few days

  • Tens of billions of objects, each one observed ∼1000 times

SDSS

Image credit: Peter Melchior

DES

Image credit: Peter Melchior

Image credit: Peter Melchior

HSC (proxy for LSST)

The Deep Learning Boom in Astrophysics

astro-ph abstracts mentioning Deep Learning, CNN, or Neural Networks

The vast majority of these results has relied on supervised learning and networks trained from scratch.

The Limits of Traditional Deep Learning

  • Limited Supervised Training Data
    • ​Rare or novel objects have by definition few labeled examples
       
    • In Simulation Based Inference (SBI), training a neural compression model requires many simulations
       
  • Limited Reusability
    • Existing models are trained supervised on a specific task, and specific data.

Huang et al. (2019)

Zhang, Bloom, Gaudi, Lanusse, Lam, Lu (2021)

=> Limits in practice the ease of using deep learning for analysis and discovery

Can we make use of all the unlabeled data we have access to?

The Rise of The Foundation Model Paradigm

  • Foundation Model approach
    • Pretrain models on pretext tasks, without supervision, on very large scale datasets.
    • Adapt pretrained models to downstream tasks. 
    • Combine pretrained modules in more complex systems.

Bommasani et al. 2021

The Advantage of Scale of Data and Compute

Liu et al. 2022

Linearly Accessible Information

Zhai et al. 2022

  • Backbone of modern architectures embed input images  as vectors in         where d can typically be between 512 to 2048.
     
  • Linear probing refers to training a single matrix to adapt this vector representation to the desired downstream task.
\mathbb{R}^{d}
y = \mathbf{M} \ h_\theta(x)

Rethinking the way we use Deep Learning 

Conventional scientific workflow with deep learning

  • Build a large training set of realistic data
  • Design a neural network architecture for your data 
  • Deal with data preprocessing/normalization issues
  • Train your network on some GPUs for a day or so
  • Apply your network to your problem
  • Throw the network away...
    => Because it's completely specific to your data, and to the one task it's trained for.

Conventional researchers @ CMU
 Circa 2016

CMU DeepLens (Lanusse et al 2017)

Rethinking the way we use Deep Learning 

Foundation Model-based Scientific Workflow

  • Build a small training set of realistic data
  • Design a neural network architecture for your data
  • Deal with data preprocessing/normalization issues
  • Adapt a model in a matter of minutes
  • Apply your model to your problem
  • Throw the network away...
    => Because it's completely specific to your data, and to the one task it's trained for.

Bommasani et al. 2021

Already taken care of

y = \mathbf{M} \ h_\theta(x)

What This New Paradigm Could Mean for Us

  • Never have to retrain my own neural networks from scratch
    • Existing pre-trained models would already be near optimal, no matter the task at hand
    • Saves a lot of time and energy
       
  • Practical large scale Deep Learning even in very few example regime
    • Searching for very rare objects in large surveys like Euclid or LSST becomes possible
    • Pretraining on data itself ensures that all sorts of image artifacts are already folded in the training.
       
  • If the information is embedded in a space where it becomes linearly accessible,  very simple analysis tools are enough for downstream analysis
    • In the future, survey pipelines may add vector embedding of detected objects into catalogs, these would be enough for most tasks, without the need to go back to pixels

... but how does it work?

Masked Modeling

Sometimes, the simplest things just work...

An idea that comes from language models

p( x_m | x_0, \ldots, x_{m-1}, x_{m+1}, \ldots, x_{N})
p(x_{i+1}| x_{0},\ldots, x_{i})

i.e. BERT (Devlin et al. 2018)

Masked Autoencoders Are Scalable Vision Learners (He et al. 2021)

Masked Auto Encoding (MAE)

\mathcal{L} = \parallel M \odot ( x - f_\theta(x)) \parallel_2^2

How to use such a model for classification

Credit: (Liang et al. 2022)

Application to Dense Prediction Tasks

Kirillov et al. 2023

Example of Application on Astronomical Data: Specformer

  • Pre-Training Galaxy Spectra Representation by Masked Modeling
\mathcal{L}_{\textrm{MM}} = \frac{1}{NK} \sum_{j=1}^K \sum_{i=1}^N \textbf{m}_i \cdot (\textbf{x}_{i} - \hat{\textbf{x}}_{i})^2,

Multi-View Self-Supervised Contrastive Learning

MultiView Contrastive Learning e.g. SimCLR (Chen et al. 2000)

Contrastive Learning in Astrophysics

Self-Supervised similarity search for large scientific datasets (Stein et al. 2021)

https://georgestein-galaxy-search-streamlit-app-o7accj.streamlit.app/

Detecting Galaxy Tidal Features Using Self-Supervised Representation Learning

Project led by Alice Desmons, Francois Lanusse, Sarah Brough

Credit: Piotr bojanowski

DiNOv2 (Oquab et al. 2023) A State of the Art Model

PCA of patch features

Dense Semantic Segmentation

Dense Depth Estimation

Weakly Supervised Contrastive Learning

Or what you can do when you do have independent views of an object...

Contrastive Language Image Pretraining

Contrastive Language Image Pretraining (CLIP)
(Radford et al. 2021)

The Information Point of View

  • The InfoNCE loss is a lower bound on the Mutual Information between modalities

Shared  information 

van den Oord et al. (2018)

(Oquab et al. 2023)

One model, many downstream applications!

Flamingo: a Visual Language Model for Few-Shot Learning (Alayrac et al. 2022)

Hierarchical Text-Conditional Image Generation with CLIP Latents (Ramesh et al. 2022)

AstroCLIP

Cross-Modal Pre-Training for Astronomical Foundation Models

Project led by Francois Lanusse, Liam Parker, Siavash Golkar, Miles Cranmer
Accepted contribution at the NeurIPS 2023 AI4Science Workshop

The AstroCLIP approach

  • We use spectra and multi-band images as our two different views for the same underlying object.
     
  • DESI Legacy Surveys (g,r,z) images, and DESI EDR galaxy spectra.

Cosine similarity search

The AstroCLIP Model (v2, Parker et al. in prep.)

  • For images, we use a ViT-L Transformer.
     
  • For spectra, we use a decoder only Transformer working at the level of spectral patches.

(Dosovitskiy et al 2021)

Evaluation of the model

  • Cross-Modal similarity search

Image Similarity

Spectral Similarity

Image-Spectral Similarity

  • Redshift Estimation From Images

Supervised baseline

z_{true}
z_{true}
z_{true}
z_{true}
z_{true}
z_{true}
z_{true}
z_{true}
  • Zero-shot prediction                    
    • k-NN regression

 

 

  • Few-shot prediction
    • MLP head trained on top of frozen backbone
  • Galaxy Physical Property Estimation from Images and Spectra

We use estimates of galaxy properties from the PROVABGS catalog (Hahn et al. 2023) (Bayesian spectral energy distribution (SED) modeling of DESI spectroscopy and photometry method)

R^2

of regression

Negative Log Likelihood of Neural Posterior Inference

  • Galaxy Morphology Classification

Classification Accuracy

We test a galaxy morphology classification task using as labels the GZ-5 dataset (Walmsley et al. 2021)

Now you try it!

https://tinyurl.com/astroclip-tuto

https://tinyurl.com/astroclip-tuto-solution

Multimodal Masked Modeling

Or let's do one massive model of everything!

AION-1
AstronomIcal Omnimodal Network

with extensive  support from the rest of the team.

Project led by:

Francois
Lanusse

Liam
Parker

Jeff
Shen

Tom
  Hehir

Ollie
Liu

Lucas
Meyer

Leopoldo
Sarra

Sebastian Wagner-Carena

Helen
Qu

Micah
Bowles

Diverse data modalities for diverse science cases

(Blanco Telescope and Dark Energy Camera.
Credit: Reidar Hahn/Fermi National Accelerator Laboratory)

(Subaru Telescope and Hyper Suprime Cam. Credit: NAOJ)

(Dark Energy Spectroscopic Instrument)

(Sloan Digital Sky Survey. Credit: SDSS)

(Gaia Satellite. Credit: ESA/ATG)

  • Galaxy formation
  • Cosmology
  • Stellar physics
  • Galaxy archaeology
  • ...

 

Standardizing all modalities through tokenization 

  • For each modality class (e.g. image, spectrum) we build dedicated metadata-aware tokenizers
     
  • For Aion-1, we integrate 39 different modalities (different instruments, different measurements, etc.)
\mathcal{L} = \parallel \Sigma^{- \frac{1}{2}} \left( x - d_\theta( \lfloor e_\theta(x) \rfloor_{\text{FSQ}} \right) \parallel_2^2

Field Embedding Strategy Developed for
Multiple Physics Pretraining (McCabe et al. 2023)

DES g

DES r

DES i

DES z

HSC g

HSC r

HSC i

HSC z

HSC y

(Mentzer et al. 2023)

Any-to-Any Modeling with Generative Masked Modeling

  • Training is done by pairing observations of the same objects from different instruments.
  • Each input token is tagged with a modality embedding that specifies provenance metadata.
     
  • Model is trained by cross-modal generative masked modeling (Mizrahi et al. 2023)
    => Learns the joint and all conditional distributions of provided modalities:  
\forall m,n \quad p(x_m | x_n)

AION-1 family of models

  • Models trained as part of the 2024 Jean Zay Grand Challenge, following an extension to a new partition of 1400 H100s
  • AION-1 Base: 300 M parameters
    • 64 H100s - 1.5 days
  • AION-1 Large: 800 M parameters
    • 100 H100s - 2.5 days
  • AION-1 XLarge: 3B parameters 
    • 288 H100s - 3.5 days

Adaptation of AION-1 embeddings

Adaptation at low cost 
with simple strategies:

 

  • Mean pooling + linear probing


     
  • Attentive pooling
y = \mathbf{M} \sum_i z_i
y = \operatorname{softmax} \left(\frac{\mathbf{Q} \mathbf{K}^\top(z)}{\sqrt{d}} \right) \mathbf{V}(z)
  • Can be used trivially on any input data AION-1 was trained for
     
  • Flexible to varying number/types of inputs
    => Allows for trivial data fusion

Morphology classification by Linear Probing

Trained on ->

Eval on ->

Physical parameter estimation and data fusion

Inputs:

measured fluxes

Inputs:

measured fluxes + image

Semantic segmentation

Segmenting central bar and spiral arms in galaxy images based on Galaxy Zoo 3D

(Masters et al. 2021)

Example-based retrieval from mean pooling embeddings