An Unsupervised Hunt for Gravitational Lenses

Stephen Sheng, Keerthi Vasan G C, Chi Po P Choi, James Sharpnack, Tucker Jones
Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, PMLR 151:9827-9843, 2022.

Abstract

Strong gravitational lenses allow us to peer into the farthest reaches of space by bending the light from a background object around a massive object in the foreground. Unfortunately, these lenses are extremely rare, and manually finding them in astronomy surveys is difficult and time-consuming. We are thus tasked with finding them in an automated fashion with few, if any, known lenses to form positive samples. To assist us with training, we can simulate realistic lenses within our survey images to form positive samples. Naively training a ResNet model with these simulated lenses results in a poor precision for the desired high recall, because the simulations contain artifacts that are learned by the model. In this work, we develop a lens detection method that combines simulation, data augmentation, semi-supervised learning, and GANs to improve this performance by an order of magnitude. We perform ablation studies and examine how performance scales with the number of non-lenses and simulated lenses. These findings allow researchers to go into a survey mostly "blind" and still be able to classify strong gravitational lenses with high precision and recall.

Cite this Paper

BibTeX
@InProceedings{pmlr-v151-sheng22a, title = { An Unsupervised Hunt for Gravitational Lenses }, author = {Sheng, Stephen and Vasan G C, Keerthi and Po P Choi, Chi and Sharpnack, James and Jones, Tucker}, booktitle = {Proceedings of The 25th International Conference on Artificial Intelligence and Statistics}, pages = {9827--9843}, year = {2022}, editor = {Camps-Valls, Gustau and Ruiz, Francisco J. R. and Valera, Isabel}, volume = {151}, series = {Proceedings of Machine Learning Research}, month = {28--30 Mar}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v151/sheng22a/sheng22a.pdf}, url = {https://proceedings.mlr.press/v151/sheng22a.html}, abstract = { Strong gravitational lenses allow us to peer into the farthest reaches of space by bending the light from a background object around a massive object in the foreground. Unfortunately, these lenses are extremely rare, and manually finding them in astronomy surveys is difficult and time-consuming. We are thus tasked with finding them in an automated fashion with few, if any, known lenses to form positive samples. To assist us with training, we can simulate realistic lenses within our survey images to form positive samples. Naively training a ResNet model with these simulated lenses results in a poor precision for the desired high recall, because the simulations contain artifacts that are learned by the model. In this work, we develop a lens detection method that combines simulation, data augmentation, semi-supervised learning, and GANs to improve this performance by an order of magnitude. We perform ablation studies and examine how performance scales with the number of non-lenses and simulated lenses. These findings allow researchers to go into a survey mostly "blind" and still be able to classify strong gravitational lenses with high precision and recall. } }
Endnote
%0 Conference Paper %T An Unsupervised Hunt for Gravitational Lenses %A Stephen Sheng %A Keerthi Vasan G C %A Chi Po P Choi %A James Sharpnack %A Tucker Jones %B Proceedings of The 25th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2022 %E Gustau Camps-Valls %E Francisco J. R. Ruiz %E Isabel Valera %F pmlr-v151-sheng22a %I PMLR %P 9827--9843 %U https://proceedings.mlr.press/v151/sheng22a.html %V 151 %X Strong gravitational lenses allow us to peer into the farthest reaches of space by bending the light from a background object around a massive object in the foreground. Unfortunately, these lenses are extremely rare, and manually finding them in astronomy surveys is difficult and time-consuming. We are thus tasked with finding them in an automated fashion with few, if any, known lenses to form positive samples. To assist us with training, we can simulate realistic lenses within our survey images to form positive samples. Naively training a ResNet model with these simulated lenses results in a poor precision for the desired high recall, because the simulations contain artifacts that are learned by the model. In this work, we develop a lens detection method that combines simulation, data augmentation, semi-supervised learning, and GANs to improve this performance by an order of magnitude. We perform ablation studies and examine how performance scales with the number of non-lenses and simulated lenses. These findings allow researchers to go into a survey mostly "blind" and still be able to classify strong gravitational lenses with high precision and recall.
APA
Sheng, S., Vasan G C, K., Po P Choi, C., Sharpnack, J. & Jones, T.. (2022). An Unsupervised Hunt for Gravitational Lenses . Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 151:9827-9843 Available from https://proceedings.mlr.press/v151/sheng22a.html.

Related Material