The first analytical expression to estimate photometric redshifts suggested by a machine

Krone-Martins, A.; Ishida, Emille E. O.

doi:10.1093/mnrasl/slu067

Cited by 25 publications

(16 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…models [22], template fitting methods [23], and other schemes and analyses [24], [25]. Feature selection schemes have been successfully applied as well in this context [26], [27], [28] and, in particular, using linear models for photometric redshift estimation [29]. We are, however, not aware of any approaches conducting exact feature selection for linear regression in this context.…”

Section: Applicationmentioning

confidence: 99%

Massively-parallel best subset selection for ordinary least-squares regression

Gieseke

Polsterer

Mahabal

et al. 2017

2017 IEEE Symposium Series on Computational Intelligence (SSCI)

View full text Add to dashboard Cite

Selecting an optimal subset of k k k out of d d d features for linear regression models given n n n training instances is often considered intractable for feature spaces with hundreds or thousands of dimensions. We propose an efficient massivelyparallel implementation for selecting such optimal feature subsets in a brute-force fashion for small k k k. By exploiting the enormous compute power provided by modern parallel devices such as graphics processing units, it can deal with thousands of input dimensions even using standard commodity hardware only. We evaluate the practical runtime using artificial datasets and sketch the applicability of our framework in the context of astronomy.

show abstract

Section: Applicationmentioning

confidence: 99%

Massively-parallel best subset selection for ordinary least-squares regression

Gieseke

Polsterer

Mahabal

et al. 2017

2017 IEEE Symposium Series on Computational Intelligence (SSCI)

View full text Add to dashboard Cite

show abstract

“…Initially the training set was used to map a polynomial function between the colors and the redshift (e.g., Connolly et al, 1995;Brunner et al, 1997). More recently, this process has been extended to machine learning algorithms, including artificial neural networks (e.g., Collister & Lahav, 2004;Oyaizu et al, 2008b;Bonnett, 2013), boosted decision trees (e.g., Gerdes et al, 2010), random forest (e.g., Carliles et al, 2010;Carrasco Kind & Brunner, 2013c), nearest neighbors (e.g., Ball et al, 2007Ball et al, , 2008Lima et al, 2008), spectral connectivity analysis (e.g., Freeman et al, 2009), Gaussian process (e.g., Way et al, 2009;Bonfield et al, 2010), support vector machines (e.g., Wadadekar, 2005), Quasi Newton Algorithm (e.g., Cavuoti et al, 2012;Brescia et al, 2014), and from analytical forms suggested by computational algorithms (e.g., Schmidt & Lipson, 2009;Krone-Martins et al, 2014). While only a few of these photo-z methods are publicly available, they all perform to a similar accuracy and provide only a single redshift estimate rather than a full redshift probability density function for each galaxy.…”

Section: Machine Learning Techniquesmentioning

confidence: 99%

Probabilistic Photometric Redshifts in the Era of Petascale Astronomy

Kind¹

2014

View full text Add to dashboard Cite

This thesis would certainly not have come to its completion without the help, support and trust of colleagues, friends and family. First, I would like to sincerely thank my advisor Robert J. Brunner for his help, guidance, advice, support and freedom during this work, for believing in me even before starting the thesis and for our extended brainstormings and interesting conversations about research, science, computing, family life and obviously soccer and for making this thesis a enjoyable experience. I am also grateful to the members of the thesis committee for their interest, comments and for taking the time to evaluate my work.I also want to thank my friends and colleagues in the Astronomy department who contributed to create and maintain a friendly and stimulating working environment during these four years. To Gary for his friendship, his help, our conversations and for suffering with me through several classes during our coursework.To Rukmani for her friendship, support and for her coffee companion during these years and for making CosmoCoffee a real thing. To Manuel and David for our lunches and lively Spanish discussions and friendship.To Jessica, Dom, Nachi, Andy, Michael, Margaret, Ian, Ashley, Ricky, and many others for being very friendly, for their understanding that I wish I'd had more time to spend with them. Special thanks Yiran and Edward for their support, discussions and for making our office a very enjoyable and friendly place. I would also like to thank many other people in this department for their support, advice and help: Special thanks to Judy and in particular to Jeri for their substantial help and for making this department a home-like place to work.I also want to acknowledge and thank my parents for their support from such a long distance, for doing the best in providing me with the best education and tools that helped to become the person I am today.Without their hard work, love, and support, none of this would have been possible. To my sisters for their love, support and for always being there for us. I would also like to sincerely thank my parents-in-law and sister-in-law who constantly traveled and helped us when we needed them. Their support and company v during the born and care of our three kids is unimaginable and without their help I couldn't have gotten through this thesis.Last but definitely not least I would like to express my infinite gratitude to my wife Andrea, to whom this work is dedicated. Without her love, company, support, help, kindness, and advice I would not have been able to do this thesis. During these four years there was not a moment in which she wasn't supporting and encouraging me, especially on the toughest and stressful of times she was able to remain calm and to provide love and peace to overcome them. All of this while being responsible for the care of our three happy and active little monsters. For her admirable dedication, love and unconditional support I will be eternally grateful. Technical acknowledgmentsI gratefully acknowledge the use of the para...

show abstract

“…Crisci et al 2012;Libbrecht & Noble 2015;Vidyasagar 2015). Following this trend, the present work is an additional effort to popularize modern machine learning techniques within astronomy (see Ball & Brunner 2010;Krone-Martins et al 2014;Ivezic et al 2014, and references therein).…”

mentioning

confidence: 93%

Exploring the spectroscopic diversity of Type Ia supernovae with dracula: a machine learning approach

Sasdelli

Ishida

Vilalta

et al. 2016

Mon. Not. R. Astron. Soc.

Self Cite

View full text Add to dashboard Cite

The existence of multiple subclasses of type Ia supernovae (SNeIa) has been the subject of great debate in the last decade. One major challenge inevitably met when trying to infer the existence of one or more subclasses is the time consuming, and subjective, process of subclass definition. In this work, we show how machine learning tools facilitate identification of subtypes of SNe Ia through the establishment of a hierarchical group structure in the continuous space of spectral diversity formed by these objects. Using Deep Learning, we were capable of performing such identification in a 4 dimensional feature space (+1 for time evolution), while the standard Principal Component Analysis barely achieves similar results using 15 principal components. This is evidence that the progenitor system and the explosion mechanism can be described by a small number of initial physical parameters. As a proof of concept, we show that our results are in close agreement with a previously suggested classification scheme and that our proposed method can grasp the main spectral features behind the definition of such subtypes. This allows the confirmation of the velocity of lines as a first order effect in the determination of SN Ia subtypes, followed by 91bg-like events. Given the expected data deluge in the forthcoming years, our proposed approach is essential to allow a quick and statistically coherent identification of SNeIa subtypes (and outliers). All tools used in this work were made publicly available in the Python package DRACULA (Dimensionality Reduction And Clustering for Unsupervised Learning in Astronomy) and can be found within COINtoolbox (https://github.com/COINtoolbox/DRACULA).

show abstract

The first analytical expression to estimate photometric redshifts suggested by a machine

Cited by 25 publications

References 31 publications

Massively-parallel best subset selection for ordinary least-squares regression

Massively-parallel best subset selection for ordinary least-squares regression

Probabilistic Photometric Redshifts in the Era of Petascale Astronomy

Exploring the spectroscopic diversity of Type Ia supernovae with dracula: a machine learning approach

Contact Info

Product

Resources

About