Differentiable strong lensing: uniting gravity and neural nets through differentiable probabilistic programming

@article{Chianese2019DifferentiableSL,
  title={Differentiable strong lensing: uniting gravity and neural nets through differentiable probabilistic programming},
  author={Marco Chianese and Adam Coogan and Paul Hofma and Sydney Otten and Christoph Weniger},
  journal={Monthly Notices of the Royal Astronomical Society},
  year={2019}
}
Since upcoming telescopes will observe thousands of strong lensing systems, creating fully automated analysis pipelines for these images becomes increasingly important. In this work, we make a step towards that direction by developing the first end-to-end differentiable strong lensing pipeline. Our approach leverages and combines three important computer science developments: (i) convolutional neural networks (CNNs), (ii) efficient gradient-based sampling techniques, and (iii) deep… 
View PDF on arXiv
22 Citations
Background Citations
10
Methods Citations
4

Figures from this paper

22 Citations

Quantifying the structure of strong gravitational lens potentials with uncertainty-aware deep neural networks

A deep learning approach to quantify the statistical properties of such perturbations directly from images, where only the extended lensed source features within a mask are considered, without the need of any lens modelling.

Differentiable probabilistic programming for strong gravitational lensing

  • M. Chianese
  • Physics
    Proceedings of 36th International Cosmic Ray Conference — PoS(ICRC2019)
  • 2019
The difficult task of observing Dark Matter subhaloes is of paramount importance since it would constrain Dark Matter particle properties (cold or warm relic) and confirm once again the longstanding…

Targeted Likelihood-Free Inference of Dark Matter Substructure in Strongly-Lensed Galaxies

A new analysis pipeline is presented that tackles diverse challenges in optical images of galaxy-galaxy strong gravitational lensing systems by bringing together many recent machine learning developments in one coherent approach, including variational inference, Gaussian processes, differentiable probabilistic programming, and neural likelihood-to-evidence ratio estimation.

Kernel Phase and Coronagraphy with Automatic Differentiation

A new Python package morphine is presented that incorporates existing kernel phase theory within this framework and an extension to the case of a Lyot coronagraph, which is found to have self-calibrating combinations of speckles.

Dark Matter Subhalos, Strong Lensing and Machine Learning

This work uses convolutional neural networks to perform a multi-classification sorting of lensed images and sees that the algorithm is able to correctly identify the lower mass cut-off within an order of magnitude to better than 93% accuracy.

The Dawes Review 10: The impact of deep learning for the analysis of galaxy surveys

This review shows that the majority of works using deep learning up to date are oriented to computer vision tasks, however, the applications are becoming more diverse and deep learning is used for estimating galaxy properties, identifying outliers or constraining the cosmological model.

Scalable Bayesian Inference for Finding Strong Gravitational Lenses

Finding strong gravitational lenses in astronomical images allows us to assess cosmological theories and understand the large-scale structure of the universe. Previous works on lens detection do not…

Strong-Lensing Source Reconstruction with Denoising Diffusion Restoration Models

Analysis of galaxy–galaxy strong lensing systems is strongly dependent on any prior assumptions made about the appearance of the source. Here we present a method of imposing a data-driven prior /…

Fast and robust Bayesian Inference using Gaussian Processes with GPry

The GPry algorithm for fast Bayesian inference of general (non-Gaussian) posteriors with a moderate number of parameters is presented, based on generating a Gaussian Process surrogate model of the log-posterior aided by a Support Vector Machine classifier that excludes extreme or non-flnite values.

Strong Lensing Source Reconstruction Using Continuous Neural Fields

From the nature of dark matter to the rate of expansion of our Universe, observations of distant galaxies distorted through strong gravitational lensing have the potential to answer some of the major…

References

SHOWING 1-10 OF 114 REFERENCES

Fast automated analysis of strong gravitational lenses with convolutional neural networks

The use of deep convolutional neural networks are reported to be used to estimate lensing parameters in an extremely fast and automated way, circumventing the difficulties that are faced by maximum likelihood methods.

Uncertainties in Parameters Estimated with Neural Networks: Application to Strong Gravitational Lensing

The results suggest that the application of approximate Bayesian neural networks to astrophysical modeling problems can be a fast alternative to Monte Carlo Markov Chains, allowing orders of magnitude improvement in speed.

Direct detection of dark matter substructure in strong lens images with convolutional neural networks

Strong gravitational lensing is a promising way of uncovering the nature of dark matter, by finding perturbations to images that cannot be well accounted for by modeling the lens galaxy without…

AutoLens: automated modeling of a strong lens’s light, mass, and source

The intricate analysis of a strong gravitational lens is a complex and computationally demanding problem, requiring the lensed source galaxy's extended light profile to be reconstructed…

THE THIRD GRAVITATIONAL LENSING ACCURACY TESTING (GREAT3) CHALLENGE HANDBOOK

The GRavitational lEnsing Accuracy Testing 3 (GREAT3) challenge is the third in a series of image analysis challenges, with a goal of testing and facilitating the development of methods for analyzing…

CosmoGAN: creating high-fidelity weak lensing convergence maps using Generative Adversarial Networks

This work applies Generative Adversarial Networks to the problem of generating weak lensing convergence maps and shows that the generator network produces maps that are described by, with high statistical confidence, the same summary statistics as the fully simulated maps.

Mining for Dark Matter Substructure: Inferring Subhalo Population Properties from Strong Lenses with Machine Learning

The subtle and unique imprint of dark matter substructure on extended arcs in strong-lensing systems contains a wealth of information about the properties and distribution of dark matter on small…

Deep Learning the Morphology of Dark Matter Substructure

This work implements, for the first time, a classification approach to identifying dark matter based on simulated strong lensing images with different substructure based on convolutional neural networks trained on sets of simulated images.

A Bayesian analysis of regularized source inversions in gravitational lensing

Strong gravitational lens systems with extended sources are of special interest because they provide additional constraints on the models of the lens systems. To use a gravitational lens system for…

Deblending galaxy superpositions with branched generative adversarial networks

It is shown that generative models are a powerful engine for deblending given their innate ability to infill missing pixel values occluded by the superposition, and a novel branched generative adversarial network to deblend overlapping galaxies is proposed.
...