An official website of the United States government
ABSTRACT Gravitational time delays provide a powerful one-step measurement of H0, independent of all other probes. One key ingredient in time-delay cosmography are high-accuracy lens models. Those are currently expensive to obtain, both, in terms of computing and investigator time (105–106 CPU hours and ∼0.5–1 yr, respectively). Major improvements in modelling speed are therefore necessary to exploit the large number of lenses that are forecast to be discovered over the current decade. In order to bypass this roadblock, we develop an automated modelling pipeline and apply it to a sample of 31 lens systems, observed by the Hubble Space Telescope in multiple bands. Our automated pipeline can derive models for 30/31 lenses with few hours of human time and <100 CPU hours of computing time for a typical system. For each lens, we provide measurements of key parameters and predictions of magnification as well as time delays for the multiple images. We characterize the cosmography-readiness of our models using the stability of differences in the Fermat potential (proportional to time delay) with respect to modelling choices. We find that for 10/30 lenses, our models are cosmography or nearly cosmography grade (<3 per cent and 3–5 per cent variations). For 6/30 lenses, the models are close to cosmography grade (5–10 per cent). These results utilize informative priors and will need to be confirmed by further analysis. However, they are also likely to improve by extending the pipeline modelling sequence and options. In conclusion, we show that uniform cosmography grade modelling of large strong lens samples is within reach.
more »
« less
Optimizing machine learning methods to discover strong gravitational lenses in the deep lens survey
ABSTRACT Machine learning models can greatly improve the search for strong gravitational lenses in imaging surveys by reducing the amount of human inspection required. In this work, we test the performance of supervised, semi-supervised, and unsupervised learning algorithms trained with the ResNetV2 neural network architecture on their ability to efficiently find strong gravitational lenses in the Deep Lens Survey (DLS). We use galaxy images from the survey, combined with simulated lensed sources, as labeled data in our training data sets. We find that models using semi-supervised learning along with data augmentations (transformations applied to an image during training, e.g. rotation) and Generative Adversarial Network (GAN) generated images yield the best performance. They offer 5 – 10 times better precision across all recall values compared to supervised algorithms. Applying the best performing models to the full 20 deg2 DLS survey, we find 3 Grade-A lens candidates within the top 17 image predictions from the model. This increases to 9 Grade-A and 13 Grade-B candidates when 1 per cent (∼2500 images) of the model predictions are visually inspected. This is ≳ 10 × the sky density of lens candidates compared to current shallower wide-area surveys (such as the Dark Energy Survey), indicating a trove of lenses awaiting discovery in upcoming deeper all-sky surveys. These results suggest that pipelines tasked with finding strong lens systems can be highly efficient, minimizing human effort. We additionally report spectroscopic confirmation of the lensing nature of two Grade-A candidates identified by our model, further validating our methods.
more »
« less
- Award ID(s):
- 2108515
- PAR ID:
- 10439790
- Publisher / Repository:
- Oxford University Press
- Date Published:
- Journal Name:
- Monthly Notices of the Royal Astronomical Society
- Volume:
- 524
- Issue:
- 4
- ISSN:
- 0035-8711
- Format(s):
- Medium: X Size: p. 5368-5390
- Size(s):
- p. 5368-5390
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Low surface brightness galaxies (LSBGs), galaxies that are fainter than the dark night sky, are famously difficult to detect. Nonetheless, studies of these galaxies are essential to improve our understanding of the formation and evolution of low-mass galaxies. In this work, we train a deep learning model using the Mask R-CNN framework on a set of simulated LSBGs inserted into images from the Dark Energy Survey (DES) Data Release 2 (DR2). This deep learning model is combined with several conventional image pre-processing steps to develop a pipeline for the detection of LSBGs. We apply this pipeline to the full DES DR2 coadd image dataset, and preliminary results show the detection of 22 large, high-quality LSBG candidates that went undetected by conventional algorithms. Furthermore, we find that the performance of our algorithm is greatly improved by including examples of false positives as an additional class during training.more » « less
-
The need for manual and detailed annotations limits the applicability of supervised deep learning algorithms in medical image analyses, specifically in the field of pathology. Semi-supervised learning (SSL) provides an effective way for leveraging unlabeled data to relieve the heavy reliance on the amount of labeled samples when training a model. Although SSL has shown good performance, the performance of recent state-of-the-art SSL methods on pathology images is still under study. The problem for selecting the most optimal data to label for SSL is not fully explored. To tackle this challenge, we propose a semi-supervised active learning framework with a region-based selection criterion. This framework iteratively selects regions for an-notation query to quickly expand the diversity and volume of the labeled set. We evaluate our framework on a grey-matter/white-matter segmentation problem using gigapixel pathology images from autopsied human brain tissues. With only 0.1% regions labeled, our proposed algorithm can reach a competitive IoU score compared to fully-supervised learning and outperform the current state-of-the-art SSL by more than 10% of IoU score and DICE coefficient.more » « less
-
Abstract Observed evolution of the total mass distribution with redshift is crucial to testing galaxy evolution theories. To measure the total mass distribution, strong gravitational lenses complement the resolved dynamical observations that are currently limited toz≲ 0.5. Here we present the lens models for a pilot sample of seven galaxy-scale lenses from theASTRO3DGalaxy Evolution with Lenses (AGEL) survey. TheAGELlenses, modeled using HST/WFC3-F140W images with Gravitational Lens Efficient Explorer (GLEE) software, have deflector redshifts in the range 0.3 <zdefl< 0.9. Assuming a power-law density profile with slopeγ, we measure the total density profile for the deflector galaxies via lens modeling. We also measure the stellar velocity dispersions (σobs) for four lenses and obtainσobsfromSDSS-BOSSfor the remaining lenses to test our lens models by comparing observed and model-predicted velocity dispersions. For the sevenAGELlenses, we measure an average density profile slope of −1.95 ± 0.09 and aγ–zrelation that does not evolve with redshift atz< 1. Although our result is consistent with some observations and simulations, it differs from other studies atz< 1 that suggest theγ–zrelation evolves with redshift. The apparent conflicts among observations and simulations may be due to a combination of (1) systematics in the lensing and dynamical modeling; (2) challenges in comparing observations with simulations; and (3) assuming a simple power law for the total mass distribution. By providing more lenses atzdefl> 0.5, theAGELsurvey will provide stronger constraints on whether the mass profiles evolve with redshift as predicted by current theoretical models.more » « less
-
null (Ed.)ABSTRACT We report upon 3 years of follow-up and confirmation of doubly imaged quasar lenses through imaging campaigns from 2016 to 2018 with the Near-Infrared Camera2 (NIRC2) on the W. M. Keck Observatory. A sample of 57 quasar lens candidates are imaged in adaptive-optics-assisted or seeing-limited K′-band observations. Out of these 57 candidates, 15 are confirmed as lenses. We form a sample of 20 lenses adding in a number of previously known lenses that were imaged with NIRC2 in 2013–14 as part of a pilot study. By modelling these 20 lenses, we obtain K′-band relative photometry and astrometry of the quasar images and the lens galaxy. We also provide the lens properties and predicted time delays to aid planning of follow-up observations necessary for various astrophysical applications, e.g. spectroscopic follow-up to obtain the deflector redshifts for the newly confirmed systems. We compare the departure of the observed flux ratios from the smooth-model predictions between doubly and quadruply imaged quasar systems. We find that the departure is consistent between these two types of lenses if the modelling uncertainty is comparable.more » « less
