ANNz: Estimating Photometric Redshifts Using Artificial Neural Networks

By Adrian Collister, Ofer Lahav. 2003

436 citations recorded by the source corpus.

We introduce ANNz, a freely available software package for photometric redshift estimation using artificial neural networks. ANNz learns the relation between photometry and redshift from an appropriate training set of galaxies for which the redshift is already known. Where a large and representative training set is available, ANNz is a highly competitive tool when compared with traditional template‐fitting methods. The ANNz package is demonstrated on the Sloan Digital Sky Survey Data Release 1, and for this particular data set the rms redshift error in the range 0≲z≲0.7 is σrms = 0.023. Nonideal conditions (spectroscopic sets that are small or brighter than the photometric set for which redshifts are required) are simulated, and the impact on the photometric redshift accuracy is assessed.

Subfields: Artificial Intelligence, Astronomy and Astrophysics, Instrumentation

View Paper: https://www.semanticscholar.org/paper/585304f26935cc6ddd63da50076593ad065556c1

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Astronomy and Astrophysics

In Instrumentation

Citations in the corpus

Constraints on intrinsic alignment contamination of weak lensing surveys using the MegaZ-LRG sample (2010, methodology): “Around 13,000 spectroscopic objects were obtained in the 2SLAQ survey, of which about 8000 were used to train a neural network (Collister & Lahav 2004) to produce photometric redshifts, leaving approximately 5100 galaxies to verify the estimates (Collister et al. 2007; Abdalla et al. 2008).”
A Galaxy Photometric Redshift Catalog for the Sloan Digital Sky Survey Data Release 6 (2007, methodology): “We follow the notation of Collister & Lahav (2004) and denote a network with Ni input nodes, Nhj nodes in hidden layer j, and No output nodes as Ni : Nh1 : Nh2 : ... : Nhm : No.”
PHAT: PHoto-z Accuracy Testing (2010, methodology): “An error bar is assigned to each photo-z via a chain rule (see Collister & Lahav 2004, for details).”
Estimating the redshift distribution of photometric galaxy samples – II. Applications and tests of a new method (2008, methodology): “Following the notation of Collister & Lahav (2004), we denote a network with k layers and Ni nodes in the i th layer as N1 : N2 : ... : Nk.”
Galaxy and Mass Assembly (GAMA): survey diagnostics and core data release (2010, methodology): “We use the Artificial Neural Network code ANNZ (Collister & Lahav 2004) with a network architecture of N:2N:2N:1, where N = 6 is the number of inputs to the network (five photometric bands and one radius together with their errors).”
Cosmological baryonic and matter densities from 600 000 SDSS luminous red galaxies with photometric redshifts (2006, methodology): “Further details on the ANNZ software may be found in Collister & Lahav (2004).”
Modelling and interpreting spectral energy distributions of galaxies with BEAGLE (2016, methodology): “…‘datadriven’ methods are based on the application of machine-learning techniques to determine the relation between galaxy colours and redshift (e.g. Collister & Lahav 2004; Hogan, Fairbairn & Seeburn 2015), while ‘clustering-based’ methods are based on the redshift evolution of spatial…”
Photometric redshift analysis in the Dark Energy Survey Science Verification data (2014, methodology): “SkyNet and ANNz2 are the only networks adding a constraint to the outputs: ANNz2 does this on a redshift bin per redshift bin basis while SkyNet uses a softmax transformation in the final layer, adding a constraint on all redshift bins simultaneously.”
Dark Energy Survey Year 1 Results: redshift distributions of the weak-lensing source galaxies (2017, methodology): “…of techniques have been developed for estimation of redshifts from broad-band fluxes (e.g. Arnouts et al. 1999; Benı́tez 2000; Bender et al. 2001; Collister & Lahav 2004; Feldmann et al. 2006; Ilbert et al. 2006; Hildebrandt et al. 2010; Carrasco Kind & Brunner 2013; Sánchez et al. 2014; Rau…”
A SUNYAEV–ZEL'DOVICH-SELECTED SAMPLE OF THE MOST MASSIVE GALAXY CLUSTERS IN THE 2500 deg2 SOUTH POLE TELESCOPE SURVEY (2011, methodology): “Redshifts for the 12 clusters with MOSAIC-II imaging were independently measured from these data using Artificial Neural Network (ANNz Collister & Lahav 2004) software and another red-sequence method (Song et al, in prep).”
EAZY: A Fast, Public Photometric Redshift Code (2008, methodology): “For example, Collister & Lahav (2004) point out that ANNz is able to account for the internal redenning of SDSS galaxies in their training and validation samples, but if the dust model of the lightcones is incorrect, the neural networks trained on them will have systematic problems matching real…”
Herschel-ATLAS: rapid evolution of dust in galaxies over the last 5 billion years (2010, methodology): “Where spectroscopic redshifts were not available we used photometric redshifts which were produced for H-ATLAS using SDSS and UKIDSS-LAS (Lawrence et al. 2007) data and the ANNz method (Collister & Lahav 2004).”
A CRITICAL ASSESSMENT OF PHOTOMETRIC REDSHIFT METHODS: A CANDELS INVESTIGATION (2013, positive): “Empirical codes include: ANNz (Collister & Lahav 2004); Multilayer Perceptron Artificial Neural Network (Vanzella et al. 2004); ArborZ (Gerdes et al. 2010); “Empirical-χ2” (Wolf 2009); “Random Forests” (Carliles et al. 2010).”
GALAXY CLUSTERS DISCOVERED WITH A SUNYAEV–ZEL'DOVICH EFFECT SURVEY (2008, methodology): “We have used ANNz (Collister & Lahav 2004) photoz’s and color-magnitude diagrams to probe for a cluster galaxy population and to provide rough redshift estimates for each of the SPT systems.”
CFHTLenS: Improving the quality of photometric redshifts with precision photometry (2011, background): “…Puschell et al. 1982; Koo 1985; Loh & Spillar 1986; Connolly et al. 1995; Koo 1999; Beńıtez 2000; Bolzonella et al. 2000; Wolf et al. 2001; Csabai et al. 2003; Collister & Lahav 2004; Ilbert et al. 2006; Hildebrandt et al. 2008, 2010; Coupon et al. 2009), represent the only practical solution.”