Deep Learning of DESI Mock Spectra to Find Damped Ly\\textbackslashalpha\ Systems. Wang, B., Zou, J., Cai, Z., Prochaska, J. X., Sun, Z., Ding, J., Font-Ribera, A., Gonzalez, A., Herrera-Alcantar, H. K., Irsic, V., Lin, X., Brooks, D., Chabanier, S., de Belsunce, R., Palanque-Delabrouille, N., Tarle, G., & Zhou, Z. arXiv:2201.00827 [astro-ph], January, 2022. arXiv: 2201.00827![Deep Learning of DESI Mock Spectra to Find Damped Ly\\textbackslashalpha\ Systems [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper abstract bibtex We have updated and applied a convolutional neural network (CNN) machine learning model to discover and characterize damped Ly\${\textbackslash}alpha\$ systems (DLAs) based on Dark Energy Spectroscopic Instrument (DESI) mock spectra. We have optimized the training process and constructed a CNN model that yields a DLA classification accuracy above 99\${\textbackslash}%\$ for spectra which have signal-to-noise (S/N) above 5 per pixel. Classification accuracy is the rate of correct classifications. This accuracy remains above 97\${\textbackslash}%\$ for lower signal-to-noise (S/N) \${\textbackslash}approx1\$ spectra. This CNN model provides estimations for redshift and HI column density with standard deviations of 0.002 and 0.17 dex for spectra with S/N above 3 per pixel. Also, this DLA finder is able to identify overlapping DLAs and sub-DLAs. Further, the impact of different DLA catalogs on the measurement of Baryon Acoustic Oscillation (BAO) is investigated. The cosmological fitting parameter result for BAO has less than \$0.61{\textbackslash}%\$ difference compared to analysis of the mock results with perfect knowledge of DLAs. This difference is lower than the statistical error for the first year estimated from the mock spectra: above \$1.7{\textbackslash}%\$. We also compared the performance of CNN and Gaussian Process (GP) model. Our improved CNN model has moderately 14\${\textbackslash}%\$ higher purity and 7\${\textbackslash}%\$ higher completeness than an older version of GP code, for S/N \${\textgreater}\$ 3. Both codes provide good DLA redshift estimates, but the GP produces a better column density estimate by \$24{\textbackslash}%\$ less standard deviation. A credible DLA catalog for DESI main survey can be provided by combining these two algorithms.
@article{wang_deep_2022,
title = {Deep {Learning} of {DESI} {Mock} {Spectra} to {Find} {Damped} {Ly}\{{\textbackslash}alpha\} {Systems}},
url = {http://arxiv.org/abs/2201.00827},
abstract = {We have updated and applied a convolutional neural network (CNN) machine learning model to discover and characterize damped Ly\${\textbackslash}alpha\$ systems (DLAs) based on Dark Energy Spectroscopic Instrument (DESI) mock spectra. We have optimized the training process and constructed a CNN model that yields a DLA classification accuracy above 99\${\textbackslash}\%\$ for spectra which have signal-to-noise (S/N) above 5 per pixel. Classification accuracy is the rate of correct classifications. This accuracy remains above 97\${\textbackslash}\%\$ for lower signal-to-noise (S/N) \${\textbackslash}approx1\$ spectra. This CNN model provides estimations for redshift and HI column density with standard deviations of 0.002 and 0.17 dex for spectra with S/N above 3 per pixel. Also, this DLA finder is able to identify overlapping DLAs and sub-DLAs. Further, the impact of different DLA catalogs on the measurement of Baryon Acoustic Oscillation (BAO) is investigated. The cosmological fitting parameter result for BAO has less than \$0.61{\textbackslash}\%\$ difference compared to analysis of the mock results with perfect knowledge of DLAs. This difference is lower than the statistical error for the first year estimated from the mock spectra: above \$1.7{\textbackslash}\%\$. We also compared the performance of CNN and Gaussian Process (GP) model. Our improved CNN model has moderately 14\${\textbackslash}\%\$ higher purity and 7\${\textbackslash}\%\$ higher completeness than an older version of GP code, for S/N \${\textgreater}\$ 3. Both codes provide good DLA redshift estimates, but the GP produces a better column density estimate by \$24{\textbackslash}\%\$ less standard deviation. A credible DLA catalog for DESI main survey can be provided by combining these two algorithms.},
urldate = {2022-01-05},
journal = {arXiv:2201.00827 [astro-ph]},
author = {Wang, Ben and Zou, Jiaqi and Cai, Zheng and Prochaska, J. Xavier and Sun, Zechang and Ding, Jiani and Font-Ribera, Andreu and Gonzalez, Alma and Herrera-Alcantar, Hiram K. and Irsic, Vid and Lin, Xiaojing and Brooks, David and Chabanier, Solène and de Belsunce, Roger and Palanque-Delabrouille, Nathalie and Tarle, Gregory and Zhou, Zhimin},
month = jan,
year = {2022},
note = {arXiv: 2201.00827},
keywords = {Astrophysics - Astrophysics of Galaxies, Astrophysics - Cosmology and Nongalactic Astrophysics},
}
Downloads: 0
{"_id":"4Dpi6MnSK623RvdbS","bibbaseid":"wang-zou-cai-prochaska-sun-ding-fontribera-gonzalez-etal-deeplearningofdesimockspectratofinddampedlytextbackslashalphasystems-2022","author_short":["Wang, B.","Zou, J.","Cai, Z.","Prochaska, J. X.","Sun, Z.","Ding, J.","Font-Ribera, A.","Gonzalez, A.","Herrera-Alcantar, H. K.","Irsic, V.","Lin, X.","Brooks, D.","Chabanier, S.","de Belsunce, R.","Palanque-Delabrouille, N.","Tarle, G.","Zhou, Z."],"bibdata":{"bibtype":"article","type":"article","title":"Deep Learning of DESI Mock Spectra to Find Damped Ly\\\\textbackslashalpha\\ Systems","url":"http://arxiv.org/abs/2201.00827","abstract":"We have updated and applied a convolutional neural network (CNN) machine learning model to discover and characterize damped Ly\\${\\textbackslash}alpha\\$ systems (DLAs) based on Dark Energy Spectroscopic Instrument (DESI) mock spectra. We have optimized the training process and constructed a CNN model that yields a DLA classification accuracy above 99\\${\\textbackslash}%\\$ for spectra which have signal-to-noise (S/N) above 5 per pixel. Classification accuracy is the rate of correct classifications. This accuracy remains above 97\\${\\textbackslash}%\\$ for lower signal-to-noise (S/N) \\${\\textbackslash}approx1\\$ spectra. This CNN model provides estimations for redshift and HI column density with standard deviations of 0.002 and 0.17 dex for spectra with S/N above 3 per pixel. Also, this DLA finder is able to identify overlapping DLAs and sub-DLAs. Further, the impact of different DLA catalogs on the measurement of Baryon Acoustic Oscillation (BAO) is investigated. The cosmological fitting parameter result for BAO has less than \\$0.61{\\textbackslash}%\\$ difference compared to analysis of the mock results with perfect knowledge of DLAs. This difference is lower than the statistical error for the first year estimated from the mock spectra: above \\$1.7{\\textbackslash}%\\$. We also compared the performance of CNN and Gaussian Process (GP) model. Our improved CNN model has moderately 14\\${\\textbackslash}%\\$ higher purity and 7\\${\\textbackslash}%\\$ higher completeness than an older version of GP code, for S/N \\${\\textgreater}\\$ 3. Both codes provide good DLA redshift estimates, but the GP produces a better column density estimate by \\$24{\\textbackslash}%\\$ less standard deviation. A credible DLA catalog for DESI main survey can be provided by combining these two algorithms.","urldate":"2022-01-05","journal":"arXiv:2201.00827 [astro-ph]","author":[{"propositions":[],"lastnames":["Wang"],"firstnames":["Ben"],"suffixes":[]},{"propositions":[],"lastnames":["Zou"],"firstnames":["Jiaqi"],"suffixes":[]},{"propositions":[],"lastnames":["Cai"],"firstnames":["Zheng"],"suffixes":[]},{"propositions":[],"lastnames":["Prochaska"],"firstnames":["J.","Xavier"],"suffixes":[]},{"propositions":[],"lastnames":["Sun"],"firstnames":["Zechang"],"suffixes":[]},{"propositions":[],"lastnames":["Ding"],"firstnames":["Jiani"],"suffixes":[]},{"propositions":[],"lastnames":["Font-Ribera"],"firstnames":["Andreu"],"suffixes":[]},{"propositions":[],"lastnames":["Gonzalez"],"firstnames":["Alma"],"suffixes":[]},{"propositions":[],"lastnames":["Herrera-Alcantar"],"firstnames":["Hiram","K."],"suffixes":[]},{"propositions":[],"lastnames":["Irsic"],"firstnames":["Vid"],"suffixes":[]},{"propositions":[],"lastnames":["Lin"],"firstnames":["Xiaojing"],"suffixes":[]},{"propositions":[],"lastnames":["Brooks"],"firstnames":["David"],"suffixes":[]},{"propositions":[],"lastnames":["Chabanier"],"firstnames":["Solène"],"suffixes":[]},{"propositions":["de"],"lastnames":["Belsunce"],"firstnames":["Roger"],"suffixes":[]},{"propositions":[],"lastnames":["Palanque-Delabrouille"],"firstnames":["Nathalie"],"suffixes":[]},{"propositions":[],"lastnames":["Tarle"],"firstnames":["Gregory"],"suffixes":[]},{"propositions":[],"lastnames":["Zhou"],"firstnames":["Zhimin"],"suffixes":[]}],"month":"January","year":"2022","note":"arXiv: 2201.00827","keywords":"Astrophysics - Astrophysics of Galaxies, Astrophysics - Cosmology and Nongalactic Astrophysics","bibtex":"@article{wang_deep_2022,\n\ttitle = {Deep {Learning} of {DESI} {Mock} {Spectra} to {Find} {Damped} {Ly}\\{{\\textbackslash}alpha\\} {Systems}},\n\turl = {http://arxiv.org/abs/2201.00827},\n\tabstract = {We have updated and applied a convolutional neural network (CNN) machine learning model to discover and characterize damped Ly\\${\\textbackslash}alpha\\$ systems (DLAs) based on Dark Energy Spectroscopic Instrument (DESI) mock spectra. We have optimized the training process and constructed a CNN model that yields a DLA classification accuracy above 99\\${\\textbackslash}\\%\\$ for spectra which have signal-to-noise (S/N) above 5 per pixel. Classification accuracy is the rate of correct classifications. This accuracy remains above 97\\${\\textbackslash}\\%\\$ for lower signal-to-noise (S/N) \\${\\textbackslash}approx1\\$ spectra. This CNN model provides estimations for redshift and HI column density with standard deviations of 0.002 and 0.17 dex for spectra with S/N above 3 per pixel. Also, this DLA finder is able to identify overlapping DLAs and sub-DLAs. Further, the impact of different DLA catalogs on the measurement of Baryon Acoustic Oscillation (BAO) is investigated. The cosmological fitting parameter result for BAO has less than \\$0.61{\\textbackslash}\\%\\$ difference compared to analysis of the mock results with perfect knowledge of DLAs. This difference is lower than the statistical error for the first year estimated from the mock spectra: above \\$1.7{\\textbackslash}\\%\\$. We also compared the performance of CNN and Gaussian Process (GP) model. Our improved CNN model has moderately 14\\${\\textbackslash}\\%\\$ higher purity and 7\\${\\textbackslash}\\%\\$ higher completeness than an older version of GP code, for S/N \\${\\textgreater}\\$ 3. Both codes provide good DLA redshift estimates, but the GP produces a better column density estimate by \\$24{\\textbackslash}\\%\\$ less standard deviation. A credible DLA catalog for DESI main survey can be provided by combining these two algorithms.},\n\turldate = {2022-01-05},\n\tjournal = {arXiv:2201.00827 [astro-ph]},\n\tauthor = {Wang, Ben and Zou, Jiaqi and Cai, Zheng and Prochaska, J. Xavier and Sun, Zechang and Ding, Jiani and Font-Ribera, Andreu and Gonzalez, Alma and Herrera-Alcantar, Hiram K. and Irsic, Vid and Lin, Xiaojing and Brooks, David and Chabanier, Solène and de Belsunce, Roger and Palanque-Delabrouille, Nathalie and Tarle, Gregory and Zhou, Zhimin},\n\tmonth = jan,\n\tyear = {2022},\n\tnote = {arXiv: 2201.00827},\n\tkeywords = {Astrophysics - Astrophysics of Galaxies, Astrophysics - Cosmology and Nongalactic Astrophysics},\n}\n\n","author_short":["Wang, B.","Zou, J.","Cai, Z.","Prochaska, J. X.","Sun, Z.","Ding, J.","Font-Ribera, A.","Gonzalez, A.","Herrera-Alcantar, H. K.","Irsic, V.","Lin, X.","Brooks, D.","Chabanier, S.","de Belsunce, R.","Palanque-Delabrouille, N.","Tarle, G.","Zhou, Z."],"key":"wang_deep_2022","id":"wang_deep_2022","bibbaseid":"wang-zou-cai-prochaska-sun-ding-fontribera-gonzalez-etal-deeplearningofdesimockspectratofinddampedlytextbackslashalphasystems-2022","role":"author","urls":{"Paper":"http://arxiv.org/abs/2201.00827"},"keyword":["Astrophysics - Astrophysics of Galaxies","Astrophysics - Cosmology and Nongalactic Astrophysics"],"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/zotero/polyphant","dataSources":["7gvjSdWrEu7z5vjjj"],"keywords":["astrophysics - astrophysics of galaxies","astrophysics - cosmology and nongalactic astrophysics"],"search_terms":["deep","learning","desi","mock","spectra","find","damped","textbackslashalpha","systems","wang","zou","cai","prochaska","sun","ding","font-ribera","gonzalez","herrera-alcantar","irsic","lin","brooks","chabanier","de belsunce","palanque-delabrouille","tarle","zhou"],"title":"Deep Learning of DESI Mock Spectra to Find Damped Ly\\\\textbackslashalpha\\ Systems","year":2022}