Evaluation of probabilistic photometric redshift estimation approaches for The Rubin Observatory Legacy Survey of Space and Time (LSST)

Schmidt, Samuel J.; Malz, Alex I.; Soo, John Y. H.; Almosallam, Ibrahim; Brescia, M.; Cavuoti, S.; Cohen‐Tanugi, J.; Connolly, Andrew J.; DeRose, J.; Freeman, Peter E.; Graham, M. L.; Iyer, Kartheik G.; Jarvis, Matt J.; Kalmbach, J. Bryce; Kovacs, E.; Lee, A.B.; Longo, Giuseppe; Morrison, C. B.; Newman, Jeffrey A.; Nourbakhsh, Erfan; Nuss, E.; Pospisil, Taylor; Hugo, Tranin,; Wechsler, Risa H.; Zhou, Rongpu; Izbicki, Rafael

doi:10.1093/mnras/staa2799

Cited by 63 publications

(91 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The crucial aspect of supervised ML methods applied to photo-z prediction is that they require a knowledge base to learn the complex relationship between broad-band photometry and distance, mainly composed by a spectroscopic redshift counterpart subsample of the photometric sources used for training, validation, and blind testing. When it is available a sufficient spectroscopic coverage of the photometric parameter space, the ML models demonstrated a high photo-z prediction accuracy, although within the limits imposed by the spectroscopic sample (Brescia et al, 2019;Euclid Collaboration et al, 2020;Schmidt et al, 2020).…”

Section: General Aspects Of the Photo-z Estimation With MLmentioning

confidence: 94%

“…For the LSST survey project, a series of scientific requirements is envisaged, aimed at avoiding the domination of the statistical background noise of the cosmological sample by any systematics, in the estimation of the photometric redshifts of several billion galaxies. In this respect, the requirements specify that the photometric redshift of any individual galaxy should have a bias below 0.003, an estimation error σ z < 0.02, and a 3σ outlier rate below 10% (Schmidt et al, 2020).…”

Section: The Leverage Of ML On Photo-z Estimationmentioning

confidence: 99%

“…These are just a few examples of the crucial role played by photometric redshifts in Astrophysics, which justifies the constant and massive proliferation of proposed solutions to optimize their accurate and reliable estimation. The strong dependence of lensing accuracy and galaxy characterization from photo-z precision, together with the required availability of a wide sample of sources, pushed many large survey projects, like LSST, KiDS, and Euclid, to perform an extensive investigation campaign dedicated to a comparison among all the most popular photo-z methods (Hildebrandt et al, 2017;Euclid Collaboration et al, 2020;Schmidt et al, 2020).…”

Section: The Leverage Of ML On Photo-z Estimationmentioning

confidence: 99%

“…Furthermore, over the last decade, within large survey projects, there was a rise in challenges dedicated to the comparison among different methods following common rules and using same data, specifically provided for these contests. The Photo-z accuracy testing (PHAT) challenge (Hildebrandt et al, 2010;Cavuoti et al, 2012) was among pioneers of such initiative, more recently followed by the Euclid and LSST challenges (Euclid Collaboration et al, 2020;Schmidt et al, 2020). In all such contests, a portion of the dataset, used for the final comparison analysis, was kept hidden from the participants in order to allow a blind test, thus providing a fair comparison in the same conditions (more examples are in Abdalla et al, 2011;de Jong et al, 2017;Bilicki et al, 2018;Tanaka et al, 2018;Amaro et al, 2019;Norris et al, 2019), often with the purpose of testing a specific case of interest.…”

Section: Is the Photo-z Point Estimate Enough?mentioning

confidence: 99%

“…Some studies (Viola et al, 2015;Mandelbaum, 2018) showed that the PDFs allow to improve the accuracy of cosmological and weak lensing measurements and to ensure a sufficient analysis of the cosmological uncertainties, from weak lensing tomography, to baryon acoustic oscillations. For this reason, most of the surveys are now producing or are planning to provide photo-z catalogs, including the PDFs and their statistics, rather than just the point estimates [see for instance KiDS (de Jong et al, 2017), Euclid (Euclid Collaboration et al, 2020), and LSST (Schmidt et al, 2020)].…”

Section: Is the Photo-z Point Estimate Enough?mentioning

confidence: 99%

See 4 more Smart Citations

Photometric Redshifts With Machine Learning, Lights and Shadows on a Complex Data Science Use Case

Brescia

Cavuoti

Razim

et al. 2021

Front. Astron. Space Sci.

View full text Add to dashboard Cite

The importance of the current role of data-driven science is constantly increasing within Astrophysics, due to the huge amount of multi-wavelength data collected every day, characterized by complex and high-volume information requiring efficient and, as much as possible, automated exploration tools. Furthermore, to accomplish main and legacy science objectives of future or incoming large and deep survey projects, such as James Webb Space Telescope (JWST), James Webb Space Telescope (LSST), and Euclid, a crucial role is played by an accurate estimation of photometric redshifts, whose knowledge would permit the detection and analysis of extended and peculiar sources by disentangling low-z from high-z sources and would contribute to solve the modern cosmological discrepancies. The recent photometric redshift data challenges, organized within several survey projects, like LSST and Euclid, pushed the exploitation of the observed multi-wavelength and multi-dimensional data or ad hoc simulated data to improve and optimize the photometric redshifts prediction and statistical characterization based on both Spectral Energy Distribution (SED) template fitting and machine learning methodologies. They also provided a new impetus in the investigation of hybrid and deep learning techniques, aimed at conjugating the positive peculiarities of different methodologies, thus optimizing the estimation accuracy and maximizing the photometric range coverage, which are particularly important in the high-z regime, where the spectroscopic ground truth is poorly available. In such a context, we summarize what was learned and proposed in more than a decade of research.

show abstract

Section: General Aspects Of the Photo-z Estimation With MLmentioning

confidence: 94%

Section: The Leverage Of ML On Photo-z Estimationmentioning

confidence: 99%

Section: The Leverage Of ML On Photo-z Estimationmentioning

confidence: 99%

Section: Is the Photo-z Point Estimate Enough?mentioning

confidence: 99%

Section: Is the Photo-z Point Estimate Enough?mentioning

confidence: 99%

See 3 more Smart Citations

Photometric Redshifts With Machine Learning, Lights and Shadows on a Complex Data Science Use Case

Brescia

Cavuoti

Razim

et al. 2021

Front. Astron. Space Sci.

View full text Add to dashboard Cite

show abstract

Cosmology with the Laser Interferometer Space Antenna

Auclair

Bacon²,

Baker³

et al. 2023

Living Rev Relativ

132

View full text Add to dashboard Cite

The Laser Interferometer Space Antenna (LISA) has two scientific objectives of cosmological focus: to probe the expansion rate of the universe, and to understand stochastic gravitational-wave backgrounds and their implications for early universe and particle physics, from the MeV to the Planck scale. However, the range of potential cosmological applications of gravitational-wave observations extends well beyond these two objectives. This publication presents a summary of the state of the art in LISA cosmology, theory and methods, and identifies new opportunities to use gravitational-wave observations by LISA to probe the universe.

show abstract

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

Huertas-Company

Lanusse²

2023

Publ. Astron. Soc. Aust.

View full text Add to dashboard Cite

The amount and complexity of data delivered by modern galaxy surveys has been steadily increasing over the past years. New facilities will soon provide imaging and spectra of hundreds of millions of galaxies. Extracting coherent scientific information from these large and multi-modal data sets remains an open issue for the community and data-driven approaches such as deep learning have rapidly emerged as a potentially powerful solution to some long lasting challenges. This enthusiasm is reflected in an unprecedented exponential growth of publications using neural networks, which have gone from a handful of works in 2015 to an average of one paper per week in 2021 in the area of galaxy surveys. Half a decade after the first published work in astronomy mentioning deep learning, and shortly before new big data sets such as Euclid and LSST start becoming available, we believe it is timely to review what has been the real impact of this new technology in the field and its potential to solve key challenges raised by the size and complexity of the new datasets. The purpose of this review is thus two-fold. We first aim at summarising, in a common document, the main applications of deep learning for galaxy surveys that have emerged so far. We then extract the major achievements and lessons learned and highlight key open questions and limitations, which in our opinion, will require particular attention in the coming years. Overall, state-of-the-art deep learning methods are rapidly adopted by the astronomical community, reflecting a democratisation of these methods. This review shows that the majority of works using deep learning up to date are oriented to computer vision tasks (e.g. classification, segmentation). This is also the domain of application where deep learning has brought the most important breakthroughs so far. However, we also report that the applications are becoming more diverse and deep learning is used for estimating galaxy properties, identifying outliers or constraining the cosmological model. Most of these works remain at the exploratory level though which could partially explain the limited impact in terms of citations. Some common challenges will most likely need to be addressed before moving to the next phase of massive deployment of deep learning in the processing of future surveys; for example, uncertainty quantification, interpretability, data labelling and domain shift issues from training with simulations, which constitutes a common practice in astronomy.

show abstract

Evaluation of probabilistic photometric redshift estimation approaches for The Rubin Observatory Legacy Survey of Space and Time (LSST)

Cited by 63 publications

References 60 publications

Photometric Redshifts With Machine Learning, Lights and Shadows on a Complex Data Science Use Case

Photometric Redshifts With Machine Learning, Lights and Shadows on a Complex Data Science Use Case

Cosmology with the Laser Interferometer Space Antenna

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

Contact Info

Product

Resources

About