# New Strong Gravitational Lenses from the DESI Legacy Imaging Surveys Data Release 9

**Authors:** C. Storfer, X. Huang, A. Gu, W. Sheu, S. Banka, A. Dey, A. Jain, J. Kwon, D. Lang, V. Lee, A. Meisner, J. Moustakas, A. D. Myers, S. Tabares-Tarquinio, E. F. Schlafly, D. J. Schlegel  
**arXiv:** [2206.02764](https://arxiv.org/abs/2206.02764)  
**Date:** 2022-06-06  
**Summary page:** https://www.summarizepaper.com/en/paper/2206.02764v1/

## Key points

- Researchers searched for strong gravitational lensing systems in the DESI Legacy Imaging Surveys Data Release 9
- They used a deep residual neural network trained on known lensing systems and candidates as well as non-lenses in the same footprint to identify potential lensing systems
- The team identified 1,895 strong lens candidates out of which 1,512 were identified for the first time
- Combining their discoveries from this work with those from Papers I and II, they have discovered a total of 3,057 strong lens candidates from the Legacy Surveys
- These findings provide valuable insights into our understanding of dark matter distribution and cosmological models as well as galaxy evolution and theories of gravity at large scales
- This work highlights the power of machine learning techniques combined with large-scale astronomical surveys to uncover new phenomena in astrophysics.

## Summary

In a recent paper, researchers conducted a search for strong gravitational lensing systems in the Dark Energy Spectroscopic Instrument (DESI) Legacy Imaging Surveys Data Release 9. This is the third paper in a series, following previous works by Huang et al. in 2020 and 2021. Together, these surveys cover approximately 19,000 deg$^2$ visible from the northern hemisphere and reach a z-band AB magnitude of approximately 22.5. To identify potential lensing systems, the researchers used a deep residual neural network trained on known lensing systems and candidates as well as non-lenses in the same footprint. After applying their trained neural networks to the survey data, they visually inspected and ranked images with probabilities above a threshold. The team identified 1,895 strong lens candidates out of which 1,512 were identified for the first time. Combining their discoveries from this work with those from Papers I and II (335 and 1,210 respectively), they have discovered a total of 3,057 strong lens candidates from the Legacy Surveys. These findings provide valuable insights into our understanding of dark matter distribution and cosmological models as well as galaxy evolution and theories of gravity at large scales. Overall, this work highlights the power of machine learning techniques combined with large-scale astronomical surveys to uncover new phenomena in astrophysics.