J/MNRAS/509/3966 Galaxy Zoo DR1&2, DR5 and DeepL version (Walmsley+, 2022)
Galaxy Zoo DECaLS Detailed visual morphology measurements from volunteers and
deep learning for 314000 galaxies.
Walmsley M., Lintott C., Geron T., Kruk S., Krawczyk C., Willett K.W.,
Bamford S., Kelvin L.S., Fortson L., Gal Y., Keel W., Masters K.L.,
Mehta V., Simmons B.D., Smethurst R., Smith L., Baeten E.M., Macmillan C.
<Mon. Not. R. Astron. Soc. 509, 3966-3988>
=2022MNRAS.509.3966W 2022MNRAS.509.3966W (SIMBAD/NED BibCode)
ADC_Keywords: Galaxies ; Morphology ; Positional data ; Redshifts ; Photometry ;
Optical ; Extinction ; Galaxies, radius
Keywords: methods: data analysis - galaxies: bar - galaxies: general -
galaxies: interactions
Abstract:
We present Galaxy Zoo DECaLS: detailed visual morphological
classifications for Dark Energy Camera Legacy Survey images of
galaxies within the SDSS DR8 footprint. Deeper DECaLS images (r = 23.6
versus r = 22.2 from SDSS) reveal spiral arms, weak bars, and tidal
features not previously visible in SDSS imaging. To best exploit the
greater depth of DECaLS images, volunteers select from a new set of
answers designed to improve our sensitivity to mergers and bars.
Galaxy Zoo volunteers provide 7.5 million individual classifications
over 314000 galaxies. 140000 galaxies receive at least 30
classifications, sufficient to accurately measure detailed morphology
like bars, and the remainder receive approximately 5. All
classifications are used to train an ensemble of Bayesian
convolutional neural networks (a state-of-the-art deep learning
method) to predict posteriors for the detailed morphology of all
314000 galaxies. We use active learning to focus our volunteer effort
on the galaxies which, if labelled, would be most informative for
training our ensemble. When measured against confident volunteer
classifications, the trained networks are approximately 99 per cent
accurate on every question. Morphology is a fundamental feature of
every galaxy; our human and machine classifications are an accurate
and detailed resource for understanding how galaxies evolve.
Description:
We present the first volunteer classifications of galaxy images
collected by the Dark Energy Camera Legacy Survey (DECaLS; Dey et al.
2019AJ....157..168D 2019AJ....157..168D). This work represents the first systematic
engagement of volunteers with low-redshift images as deep as those
provided by DECaLS, and thus represents a more reliable catalogue of
detailed morphology than has hitherto been available. Our volunteer
classifications were sourced over three separate Galaxy Zoo DECaLS
(GZD) classification campaigns, GZD-1, GZD-2, and GZD-5, which
classified galaxies first released in DECaLS Data Releases 1, 2, and
5, respectively. The key practical differences are that GZD-5 uses an
improved decision tree aimed at better identification of mergers and
weak bars, and includes galaxies with just 5 total votes as well as
galaxies with 40 or more. Across all campaigns, we collect 7496325
responses from Galaxy Zoo volunteers, recording 30 or more
classifications in at least one campaign for 139919 galaxies and fewer
(approximately 5 classifications) for an additional 173870 galaxies,
totalling 313789 classified galaxies, (i.e refer to introduction
section).
For the first time in a Galaxy Zoo data release, we also provide
automated classifications made using Bayesian deep learning (Walmsley
et al. 2020MNRAS.491.1554W 2020MNRAS.491.1554W). By using our volunteer classifications to
train a deep learning algorithm, we can make detailed classifications
for all 313789 galaxies in our target sample, providing morphology
measurements faster than would be possible than relying on volunteers
alone. Bayesian deep learning allows us to learn from uncertain
volunteer responses and to estimate the uncertainty of our
predictions. It also allows us to identify which galaxies, if
labelled, would be most informative for training our classifier
(active learning). We chose to partially focus our volunteers on such
informative galaxies, requesting 40 classifications per informative
galaxy and only 5 for the remainder. Our classifier predicts
posteriors for how volunteers would have answered all decision tree
questions.
Concerning observations, our galaxy images are created from data
collected by the DECaLS survey (Dey et al. 2019AJ....157..168D 2019AJ....157..168D).
DECaLS uses the Dark Energy Camera (DECam; Flaugher et al.
2015AJ....150..150F 2015AJ....150..150F) at the 4m Blanco telescope at Cerro Tololo
Inter-American Observatory, near La Serena, Chile. DECam has a roughly
hexagonal 3.2 deg2 field of view with a pixel scale of 0.262 arcsec
per pixel. We identify galaxies in the DECaLS imaging using the
NASA-Sloan Atlas v1.0.0 (NSA). As the NSA was derived from SDSS DR8
imaging (Aihara et al. 2011ApJS..193...29A 2011ApJS..193...29A), this data release only
includes galaxies that are within both the DECaLS and SDSS DR8
footprint.
Then, after galaxy selction (i.e see section 2.2 Selection), we
convert the measured grz fluxes into RGB images following the
methodology of Lupton et al. (2004PASP..116..133L 2004PASP..116..133L), (i.e see section
2.3 RGB image construction). The images are available on Zenodo at
https://doi.org/10.5281/ zenodo.4573248. Hereafter as explained in the
section 3 and 4 Volunteers classifications and analysis, we produce
from GZD-1&2 and GZD5 decison trees two volunteers catalogues
gzdv1-2.dat and gzdv5.dat. We also select active learning galaxies
to train our Bayesian deep learning classifier (i.e see section 5
automated classifications). Our model successfully predicts posteriors
for volunteer votes to each question as shown in the table gzdauto.dat
(i.e see section 5.2 Results to notice our predicted data quality).
Finally, the section 6 Usage explains catalogues usage contexts for
which results should be used for future studies.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
gzdv1-2.dat 805 92960 *Volunteer classifications for GZ DECaLS DR1&2
gzdv5.dat 988 253286 *Volunteer classifications for GZ DECaLS DR5
gzdauto.dat 1143 313789 *Automated classifications for GZ DECaLS DR1,
DR2 and DR5 galaxies
--------------------------------------------------------------------------------
Note on gzdv1-2.dat: Please refer to the section 3 Volunteer classifications
and 6.1 Catalogues, figure A1 Classification decision tree for these GZD DR1&2
in order to understand the question structures.
Note on gzdv5.dat: Please refer to the section 3.2 Decision trees, figure 4
Classification decision tree for GZD-5 in order to understand the question
structures.
Note on gzdauto.dat: For convenience, the catalogue includes the same set of
basic astrophysical measurements copied from the NASA Sloan Atlas (NSA).
Additional measurements can be added my crossmatching on 'iauname' with the
NSA. See (https://data.sdss.org/datamodel/files/ATLAS_DATA/ATLAS_MAJOR_VERSION
/nsa.html) for the NSA schema and the section 5 Automated classifications for
details on procedure.
--------------------------------------------------------------------------------
See also:
J/MNRAS/410/166 : Morphological types from Galaxy Zoo 1 (Lintott+, 2011)
J/MNRAS/435/2835 : Morphological types from Galaxy Zoo 2 (Willett+, 2013)
J/MNRAS/461/3663 : Galaxy Zoo 2: new classification (Hart+, 2016)
J/MNRAS/526/4768 : Galaxy Zoo DESI (Walmsley+, 2023)
https://www.github.com/mwalmsley/zoobot : Jupyter notebooks
https://www.legacysurvey.org/dr5/description/ : DR5 description and images
https://doi.org/10.5281/zenodo.4196266 : all catalogues and images data
https://www.zooniverse.org/projects/zookeeper/galaxy-zoo/ : Galaxy zoo home
Byte-by-byte Description of file: gzdv1-2.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 19 A19 --- IAUname IAU name JHHMMSS.ss+DDMMSS.s (iauname)
21- 42 F22.18 deg RAdeg Right Ascension (ra) (J2000)
44- 66 F23.20 deg DEdeg Declination (dec) (J2000)
68- 82 E15.12 --- z Heliocentric redshift (redshift)
84- 95 F12.8 mag rMag ?=- Absolute magnitude from elliptical
Petrosian fluxes in rest-frame in SDSS
r band (ABmag systems) (elpetroabsmagr)
97-113 F17.8 mag Ar Galactic extinction corrected (ABmag systems)
flux in SDSS r band (sersicnmgyr)
115-126 F12.8 arcsec Pth50 Azimuthally-averaged SDSS-style 50 per cent
Petrosian light radius r-band (petro_th50)
128-139 F12.8 arcsec Pth90 Azimuthally-averaged SDSS-style 90 per cent
Petrosian light radius r-band (petro_th90)
141-152 F12.8 arcsec Pth100 Azimuthally-averaged SDSS-style Petrosian
light radius r-band (petro_theta)
154-163 A10 --- Class Classification group only pre-active ones
(upload_group)
165-169 A5 --- Active [False] The active priority galaxies flag
(activelearningon)
171-175 A5 --- DR5 [False True] Flag indicate if the galaxy is
present in volunteers DR5 data (ingzdc)
177-179 A3 --- DR1-2 [DR1 DR2] Flag indicate if the galaxy is
present in volunteers DR1 or DR2 data
(data_release)
181-182 I2 --- NbSOF Total number of volunteers who answered
the Smooth of Featured question
(smooth-or-featured_total-votes)
184-201 F18.16 --- Smooth ? Fraction of volunteers who responded
"Smooth" to the "Smooth or Featured"
question (smooth-or-featuredsmoothfraction)
203-220 F18.16 --- Featured ? Fraction of volunteers who responded
"Featured or Disk" to the "Smooth or Featured"
question (smooth-or-featured_featured-or
-disk_fraction)
222-239 F18.16 --- Artifact ? Fraction of volunteers who responded
"Artifact or Star" to the "Smooth or
Featured" question (smooth-or-featured
artifact_fraction)
241-242 I2 --- NbHR Total number of volunteers who answered
the "How rounded is it" question
(how-rounded_total-votes)
244-261 F18.16 --- RR ? Fraction of volunteers who responded
"completely round" to the "How rounded
is it" question (how-rounded_completely
_fraction)
263-280 F18.16 --- RIB ? Fraction of volunteers who responded
"In between" to the "How rounded is it"
question (how-roundedin-betweenfraction)
282-299 F18.16 --- RCS ? Fraction of volunteers who responded
"Cigar shape" to the "How rounded is it"
question (how-roundedcigar-shapedfraction)
301-302 I2 --- NbDEO Total number of volunteers who answered
the "Disk viewed edge-on" question
(disk-edge-on_total-votes)
304-321 F18.16 --- DEOyes ? Fraction of volunteers who responded "yes"
to the "Disk viewed edge-on" question
(disk-edge-onyesfraction)
323-340 F18.16 --- DEOno ? Fraction of volunteers who responded "no"
to the "Disk viewed edge-on" question
(disk-edge-onnofraction)
342-343 I2 --- NbEOB Total number of volunteers who answered the
"edge-on-bulge shape" question
(edge-on-bulge_total-votes)
345-362 F18.16 --- EOBR ? Fraction of volunteers who responded
"Rounded" to the "edge-on-bulge shape"
question (edge-on-bulgeroundedfraction)
364-381 F18.16 --- EOBB ? Fraction of volunteers who responded
"Boxy" to the "edge-on-bulge shape" question
(edge-on-bulgeboxyfraction)
383-400 F18.16 --- EOBNo ? Fraction of volunteers who responded
"No bulge" to the "edge-on-bulge shape"
question (edge-on-bulgenonefraction)
402-403 I2 --- NbSOB Total number of volunteers who answered
the "sign of bar feature" question
(bar_total-votes)
405-422 F18.16 --- BYo ? Fraction of volunteers who responded
"Yes bar" to the "sign of bar feature"
question (baryesfraction)
424-441 F18.16 --- BNo ? Fraction of volunteers who responded
"No bar" to the "sign of bar feature" question
(barnofraction)
443-444 I2 --- NbHSA Total number of volunteers who answered
the "Has spiral arms" question (has-spiral
arms_total-votes)
446-463 F18.16 --- HSAyes ? Fraction of volunteers who responded "yes"
to the "Has spiral arms" question (has-spiral
armsyesfraction)
465-482 F18.16 --- HSAno ? Fraction of volunteers who responded "no"
to the "Has spiral arms" question (has-spiral
armsnofraction)
484-485 I2 --- NbSW Total number of volunteers who answered
the "How tightly wound spiral arms appear"
question (spiral-winding_total-votes)
487-504 F18.16 --- SWT ? Fraction of volunteers who responded "Tight"
to the "How tightly" question
(spiral-windingtightfraction)
506-523 F18.16 --- SWM ? Fraction of volunteers who responded
"Medium" to the "How tightly" question
(spiral-windingmediumfraction)
525-542 F18.16 --- SWL ? Fraction of volunteers who responded "Loose"
to the "How tightly" question
(spiral-windingloosefraction)
544-545 I2 --- NbSA Total number of volunteers who answered the
"How many spiral arms" question
(spiral-arm-count_total-votes)
547-564 F18.16 --- SAC1 ? Total number of volunteers who responded
"1 arm" to the "How many spiral arms" question
(spiral-arm-count1fraction)
566-583 F18.16 --- SAC2 ? Total number of volunteers who responded
"2 arms" to the "How many spiral arms"
question (spiral-arm-count2fraction)
585-602 F18.16 --- SAC3 ? Total number of volunteers who responded
"3 arms" to the "How many spiral arms"
question (spiral-arm-count3fraction)
604-621 F18.16 --- SAC4 ? Total number of volunteers who responded
"4 arms" to the "How many spiral arms"
question (spiral-arm-count4fraction)
623-640 F18.16 --- SAC4+ ? Total number of volunteers who responded
"More than 4 arms" to the "How many spiral
arms" question (spiral-arm-count
more-than-4fraction)
642-643 I2 --- NbBulge Total number of volunteers who answered the
"bulge size" question (bulge-size_total-votes)
645-662 F18.16 --- BSNo ? Fraction of volunteers who responded "No
bulge" to the "bulge size" question
(bulge-sizenonefraction)
664-681 F18.16 --- BSO ? Fraction of volunteers who responded
"Obvious" to the "bulge size" question
(bulge-sizeobviousfraction)
683-700 F18.16 --- BSD ? Fraction of volunteers who responded "bulge
size dominant" to the "bulge size" question
(bulge-sizedominantfraction)
702-703 I2 --- NbM ? Total number of volunteers who answered
the "Galaxy merging or tidal debris" question
(merging_total-votes)
705-722 F18.16 --- MM ? Total number of volunteers who responded
"Merging galaxy" to the "Galaxy merging or
tidal debris" question
(mergingmergerfraction)
724-741 F18.16 --- MTD ? Total number of volunteers who responded
"Tidal debris" to the "Galaxy merging or
disturbed" question (merging_tidal-debris
_fraction)
743-760 F18.16 --- MBO ? Total number of volunteers who responded
"Both" to the "Galaxy merging or
tidal debris" question (mergingbothfraction)
762-779 F18.16 --- Mno ? Total number of volunteers who responded
"Neither" to the "Galaxy merging or disturbed"
question (mergingneitherfraction)
781-799 F19.15 --- WSS ? Mean distance from center of all pixels
above double the 20th percentile
(wrongsizestatistic)
801-805 A5 --- WSW [False True] True if WSS > 161.0 our suggested
starting cut to the mean distance of all
pixels from center (wrongsizewarning)
--------------------------------------------------------------------------------
Byte-by-byte Description of file: gzdv5.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 19 A19 --- IAUname IAU name JHHMMSS.ss+DDMMSS.s (iauname)
21- 42 F22.18 deg RAdeg Right Ascension (ra) (J2000)
44- 66 E23.20 deg DEdeg Declination (dec) (J2000)
68- 82 E15.12 --- z Heliocentric redshift (redshift)
84- 94 F11.7 mag rMag ?=- Absolute magnitude from elliptical
Petrosian fluxes in rest-frame in SDSS
r band (ABmag systems) (elpetroabsmagr)
96-107 E12.9 mag Ar Galactic extinction corrected (ABmag systems)
flux in SDSS r band (sersicnmgyr)
109-120 F12.8 arcsec Pth50 Azimuthally-averaged SDSS-style 50 per cent
Petrosian light radius r-band (petro_th50)
122-133 F12.8 arcsec Pth90 Azimuthally-averaged SDSS-style 90 per cent
Petrosian light radius r-band (petro_th90)
135-145 F11.7 arcsec Pth100 Azimuthally-averaged SDSS-style Petrosian
light radius r-band (petro_theta)
147-161 A15 --- Class Classification groups (upload_group) (1)
163-167 A5 --- Active [False True] The active priority galaxies flag
(activelearningon)
169-173 A5 --- DR1-2 [False True] Flag indicate if the galaxy is
present in volunteers DR1&2 data (ingzdab)
175-206 A32 --- Pngloc Image file location in the DR5 (png_loc)
208-210 I3 --- NbSOF Total number of volunteers who answered
the Smooth of Featured question
(smooth-or-featured_total-votes)
212-231 F20.18 --- Smooth ? Fraction of volunteers who responded
"Smooth" to the "Smooth or Featured" question
(smooth-or-featuredsmoothfraction)
233-252 F20.18 --- Featured ? Fraction of volunteers who responded
"Featured or Disk" to the "Smooth or Featured"
question (smooth-or-featured
featured-or-disk_fraction)
254-274 F21.19 --- Artifact ? Fraction of volunteers who responded
"Artifact or Star" to the "Smooth or
Featured" question (smooth-or-featured
artifact_fraction)
276-278 I3 --- NbDEO Total number of volunteers who answered
the "Disk viewed edge-on" question
(disk-edge-ontotal-votes)
280-299 F20.18 --- DEOyes ? Fraction of volunteers who responded "yes"
to the "Disk viewed edge-on" question
(disk-edge-onyesfraction)
301-320 F20.18 --- DEOno ? Fraction of volunteers who responded "no"
to the "Disk viewed edge-on" question
(disk-edge-onnofraction)
322-324 I3 --- NbHSA Total number of volunteers who answered
the "Has spiral arms" question (has-spiral
arms_total-votes)
326-345 F20.18 --- HSAyes ? Fraction of volunteers who responded "yes"
to the "Has spiral arms" question (has-spiral
armsyesfraction)
347-366 F20.18 --- HSAno ? Fraction of volunteers who responded "no"
to the "Has spiral arms" question (has-spiral
armsnofraction)
368-370 I3 --- NbSOB Total number of volunteers who answered
the "sign of bar feature" question
(bar_total-votes)
372-391 F20.18 --- BS ? Fraction of volunteers who responded
"strong bar" to the "sign of bar feature"
question (barstrongfraction)
393-412 F20.18 --- BW ? Fraction of volunteers who responded
"weak bar" to the "sign of bar feature"
question (barweakfraction)
414-433 F20.18 --- BNo ? Fraction of volunteers who responded
"No bar" to the "sign of bar feature" question
(barnofraction)
435-437 I3 --- NbBulge Total number of volunteers who answered the
"bulge size" question (bulge-size_total-votes)
439-458 F20.18 --- BSD ? Fraction of volunteers who responded "bulge
size dominant" to the "bulge size" question
(bulge-sizedominantfraction)
460-479 F20.18 --- BSL ? Fraction of volunteers who responded "bulge
size large" to the "bulge size" question
(bulge-sizelargefraction)
481-500 F20.18 --- BSM ? Fraction of volunteers who responded "bulge
size moderate" to the "bulge size" question
(bulge-sizemoderatefraction)
502-521 F20.18 --- BSS ? Fraction of volunteers who responded "bulge
size small" to the "bulge size" question
(bulge-sizesmallfraction)
523-542 F20.18 --- BSNo ? Fraction of volunteers who responded "No
bulge" to the "bulge size" question
(bulge-sizenonefraction)
544-546 I3 --- NbHR Total number of volunteers who answered
the "How rounded is it" question
(how-rounded_total-votes)
548-567 F20.18 --- RR ? Fraction of volunteers who responded
"round" to the "How rounded is it" question
(how-roundedroundfraction)
569-588 F20.18 --- RIB ? Fraction of volunteers who responded
"In between" to the "How rounded is it"
question (how-roundedin-betweenfraction)
590-609 F20.18 --- RCS ? Fraction of volunteers who responded
"Cigar shape" to the "How rounded is it"
question (how-roundedcigar-shapedfraction)
611-613 I3 --- NbEOB Total number of volunteers who answered the
"edge-on-bulge shape" question
(edge-on-bulge_total-votes)
615-634 F20.18 --- EOBB ? Fraction of volunteers who responded
"Boxy" to the "edge-on-bulge shape" question
(edge-on-bulgeboxyfraction)
636-655 F20.18 --- EOBNo ? Fraction of volunteers who responded
"No bulge" to the "edge-on-bulge shape"
question (edge-on-bulgenonefraction)
657-676 F20.18 --- EOBR ? Fraction of volunteers who responded
"Rounded" to the "edge-on-bulge shape"
question (edge-on-bulgeroundedfraction)
678-680 I3 --- NbSW Total number of volunteers who answered
the "How tightly wound spiral arms appear"
question (spiral-winding_total-votes)
682-701 F20.18 --- SWT ? Fraction of volunteers who responded "Tight"
to the "How tightly" question
(spiral-windingtightfraction)
703-722 F20.18 --- SWM ? Fraction of volunteers who responded
"Medium" to the "How tightly" question
(spiral-windingmediumfraction)
724-743 F20.18 --- SWL ? Fraction of volunteers who responded "Loose"
to the "How tightly" question
(spiral-windingloosefraction)
745-747 I3 --- NbSA Total number of volunteers who answered the
"How many spiral arms" question
(spiral-arm-count_total-votes)
749-768 F20.18 --- SAC1 ? Total number of volunteers who responded
"1 arm" to the "How many spiral arms" question
(spiral-arm-count1fraction)
770-789 F20.18 --- SAC2 ? Total number of volunteers who responded
"2 arms" to the "How many spiral arms"
question (spiral-arm-count2fraction)
791-810 F20.18 --- SAC3 ? Total number of volunteers who responded
"3 arms" to the "How many spiral arms"
question (spiral-arm-count3fraction)
812-831 F20.18 --- SAC4 ? Total number of volunteers who responded
"4 arms" to the "How many spiral arms"
question (spiral-arm-count4fraction)
833-852 F20.18 --- SAC4+ ? Total number of volunteers who responded
"More than 4 arms" to the "How many spiral
arms" question
(spiral-arm-countmore-than-4fraction)
854-873 F20.18 --- SACCT ? Total number of volunteers who responded
"Can't tell" to the "How many spiral arms"
question (spiral-arm-countcant-tellfraction)
875-877 I3 --- NbM ? Total number of volunteers who answered
the "Galaxy merging or disturbed" question
(merging_total-votes)
879-898 F20.18 --- Mno ? Total number of volunteers who responded
"None" to the "Galaxy merging or disturbed"
question (mergingnonefraction)
900-919 F20.18 --- MMiD ? Total number of volunteers who responded
"Minor disturbance" to the "Galaxy merging
or disturbed" question
(mergingminor-disturbancefraction)
921-940 F20.18 --- MMaD ? Total number of volunteers who responded
"Major disturbance" to the "Galaxy merging
or disturbed" question
(mergingmajor-disturbancefraction)
942-962 F21.19 --- MM ? Total number of volunteers who responded
"Merging galaxy" to the "Galaxy merging or
disturbed" question (mergingmergerfraction)
964-982 F19.15 --- WSS ? Mean distance from center of all pixels
above double the 20th percentile
(wrongsizestatistic)
984-988 A5 --- WSW [False True] True if WSS > 161.0 our suggested
starting cut to the mean distance of all
pixels from center (wrongsizewarning)
--------------------------------------------------------------------------------
Note (1): The upload_group columns values are organized as follows:
active_baseline = 192557 galaxies are cleassified as 'active_baseline'
to be remainder ones with 5 informative volunteers
classications, we then rely more on models (i.e see
the section 3 Volunteer classifications) to classify
them
active_priority = Prioritizing 6939 galaxies from the remaining pool
of 199496 galaxies not yet uploaded. The galaxies
are identified with the groups 'active_priority'
and 'upload_group' (the galaxies identified as 'most
informative' and selected for 40 classifications).
By 'most informative', we mean the galaxies which,
if classified, would most improve the performance
of our model. We describe our method for estimating
which galaxies would be most informative in full in
Walmsley et al. (2020MNRAS.491.1554W 2020MNRAS.491.1554W)
calibration = Similarly to the targeted galaxies, 651 of these
calibration galaxies would have otherwise received
five classifications, are identified with the
'calibration' group, and should be excluded
pre_active = 51994 galaxies classified 'pre-active' which have
at least classifications were active in previous
data released 1 and 2
targeted = Early galaxies were initially uploaded row-by-row
from the NASA-Sloan Atlas, each (eventually)
receiving 40 classifications. We also uploaded two
additional subsets. For the first, 1355 galaxies
were targeted for classification to support an
external research project. Of these, 1145 would
have otherwise received five classifications.
These 1145 galaxies with additional classifications
are identified with the 'targeted' group and should
be excluded
Bayesian deep learning allows us to learn from uncertain volunteer
responses and to estimate the uncertainty of our predictions. It also
allows us to identify which galaxies, if labelled, would be most
informative for training our classifier (active learning). We chose
to partially focus our volunteers on such informative galaxies,
requesting 40 classifications per informative galaxy and only 5
for the remainder.
As explained in the section 3.1 Selecting total classifications and
the section 6 Usage, we implemented active learning, prioritizing 6939
galaxies from the remaining pool of 199496 galaxies not yet uploaded.
The galaxies are identified with the groups 'active_priority'
(the galaxies identified as 'most informative' and selected for 40
classifications) and 'active_baseline' (the remainder). For a strict
random selection, both groups should be excluded, leaving the galaxies
classified prior to the introduction of active learning.
--------------------------------------------------------------------------------
Byte-by-byte Description of file: gzdauto.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 22 F22.18 deg RAdeg Right Ascension (ra) (J2000)
24- 46 E23.20 deg DEdeg Declination (dec) (J2000)
48- 66 A19 --- IAUname Unique identifier listed in NSA v1.0.1
(iauname)
68- 79 F12.8 arcsec Pth100 Azimuthally-averaged SDSS-style Petrosian
light radius r-band (petro_theta)
81- 92 F12.8 arcsec Pth50 Azimuthally-averaged SDSS-style 50 per cent
Petrosian light radius r-band (petro_th50)
94-105 F12.8 arcsec Pth90 Azimuthally-averaged SDSS-style 90 per cent
Petrosian light radius r-band (petro_th90)
107-118 F12.8 mag rMag ?=- Absolute magnitude from elliptical
Petrosian fluxes in rest-frame in SDSS
r band (ABmag systems) (elpetroabsmagr)
120-131 E12.9 mag Ar Galactic extinction corrected (ABmag systems)
flux in SDSS r band (sersicnmgyr)
133-147 E15.12 --- z ? Heliocentric redshift (redshift)
149-159 F11.7 --- rmag ?=- Magnitude from 22.5-2.5*log(Ar) (mag_r)
161-180 F20.18 --- Smooth ? Fraction of volunteers who responded
"Smooth" to the "Smooth or Featured"
question (smooth-or-featuredsmoothfraction)
182-201 F20.18 --- Featured ? Fraction of volunteers who responded
"Featured or Disk" to the "Smooth or Featured"
question (smooth-or-featured_featured-or
-disk_fraction)
203-222 F20.18 --- Artifact ? Fraction of volunteers who responded
"Artifact or Star" to the "Smooth or
Featured" question (smooth-or-featured
artifact_fraction)
224-243 F20.18 --- DEOyes ? Fraction of volunteers who responded "yes"
to the "Disk viewed edge-on" question
(disk-edge-onyesfraction)
245-264 F20.18 --- DEOno ? Fraction of volunteers who responded "no"
to the "Disk viewed edge-on" question
(disk-edge-onnofraction)
266-285 F20.18 --- HSAyes ? Fraction of volunteers who responded "yes"
to the "Has spiral arms" question (has-spiral
armsyesfraction)
287-306 F20.18 --- HSAno ? Fraction of volunteers who responded "no"
to the "Has spiral arms" question (has-spiral
armsnofraction)
308-327 F20.18 --- BS ? Fraction of volunteers who responded
"strong bar" to the "sign of bar feature"
question (barstrongfraction)
329-348 F20.18 --- BW ? Fraction of volunteers who responded
"weak bar" to the "sign of bar feature"
question (barweakfraction)
350-369 F20.18 --- BNo ? Fraction of volunteers who responded
"No bar" to the "sign of bar feature" question
(barnofraction)
371-391 F21.19 --- BSD ? Fraction of volunteers who responded "bulge
size dominant" to the "bulge size" question
(bulge-sizedominantfraction)
393-412 F20.18 --- BSL ? Fraction of volunteers who responded "bulge
size large" to the "bulge size" question
(bulge-sizelargefraction)
414-433 F20.18 --- BSM ? Fraction of volunteers who responded "bulge
size moderate" to the "bulge size" question
(bulge-sizemoderatefraction)
435-454 F20.18 --- BSS ? Fraction of volunteers who responded "bulge
size small" to the "bulge size" question
(bulge-sizesmallfraction)
456-476 F21.19 --- BSNo ? Fraction of volunteers who responded "No
bulge" to the "bulge size" question
(bulge-sizenonefraction)
478-497 F20.18 --- RR ? Fraction of volunteers who responded
"round" to the "How rounded is it" question
(how-roundedroundfraction)
499-518 F20.18 --- RIB ? Fraction of volunteers who responded
"In between" to the "How rounded is it"
question (how-roundedin-betweenfraction)
520-540 F21.19 --- RCS ? Fraction of volunteers who responded
"Cigar shape" to the "How rounded is it"
question (how-roundedcigar-shapedfraction)
542-561 F20.18 --- EOBB ? Fraction of volunteers who responded
"Boxy" to the "edge-on-bulge shape" question
(edge-on-bulgeboxyfraction)
563-582 F20.18 --- EOBNo ? Fraction of volunteers who responded
"No bulge" to the "edge-on-bulge shape"
question (edge-on-bulgenonefraction)
584-603 F20.18 --- EOBR ? Fraction of volunteers who responded
"Rounded" to the "edge-on-bulge shape"
question (edge-on-bulgeroundedfraction)
605-624 F20.18 --- SWT ? Fraction of volunteers who responded "Tight"
to the "How tightly" question
(spiral-windingtightfraction)
626-645 F20.18 --- SWM ? Fraction of volunteers who responded
"Medium" to the "How tightly" question
(spiral-windingmediumfraction)
647-666 F20.18 --- SWL ? Fraction of volunteers who responded "Loose"
to the "How tightly" question
(spiral-windingloosefraction)
668-687 F20.18 --- SAC1 ? Total number of volunteers who responded
"1 arm" to the "How many spiral arms" question
(spiral-arm-count1fraction)
689-708 F20.18 --- SAC2 ? Total number of volunteers who responded
"2 arms" to the "How many spiral arms"
question (spiral-arm-count2fraction)
710-730 F21.19 --- SAC3 ? Total number of volunteers who responded
"3 arms" to the "How many spiral arms"
question (spiral-arm-count3fraction)
732-752 F21.19 --- SAC4 ? Total number of volunteers who responded
"4 arms" to the "How many spiral arms"
question (spiral-arm-count4fraction)
754-774 F21.19 --- SAC4+ ? Total number of volunteers who responded
"More than 4 arms" to the "How many spiral
arms" question
(spiral-arm-countmore-than-4fraction)
776-795 F20.18 --- SACCT ? Total number of volunteers who responded
"Can't tell" to the "How many spiral arms"
question (spiral-arm-countcant-tellfraction)
797-816 F20.18 --- Mno ? Total number of volunteers who responded
"None" to the "Galaxy merging or disturbed"
question (mergingnonefraction)
818-837 F20.18 --- MMiD ? Total number of volunteers who responded
"Minor disturbance" to the "Galaxy merging
or disturbed" question
(mergingminor-disturbancefraction)
839-858 F20.18 --- MMaD ? Total number of volunteers who responded
"Major disturbance" to the "Galaxy merging
or disturbed" question
(mergingmajor-disturbancefraction)
860-879 F20.18 --- MM ? Total number of volunteers who responded
"Merging galaxy" to the "Galaxy merging or
disturbed" question (mergingmergerfraction)
881-883 F3.1 --- SmoothP Estimated proportion of volunteers asked for
the "Smooth or Featured" question (smooth-or-
featuredproportionvolunteers_asked) (1)
885-904 F20.18 --- DEOP Estimated proportion of volunteers asked for
the "Disk viewed edge-on" question (disk-edge
-onproportionvolunteers_asked) (1)
906-925 F20.18 --- HSAP Estimated proportion of volunteers asked for
the "Has spiral arms" question (has-spiral-
armsproportionvolunteers_asked) (1)
927-946 F20.18 --- SOBP Estimated proportion of volunteers asked for
the "sign of bar feature" question
(barproportionvolunteers_asked) (1)
948-967 F20.18 --- BulgeP Estimated proportion of volunteers asked for
the "bulge size" question (bulge
-sizeproportionvolunteers_asked) (1)
969-988 F20.18 --- HRP Estimated proportion of volunteers asked for
the "How rounded is it" question (how
-roundedproportionvolunteers_asked) (1)
990-1010 F21.19 --- EOBP Estimated proportion of volunteers asked for
the "edge-on-bulge shape" question (edge
-on-bulgeproportionvolunteers_asked) (1)
1012-1032 F21.19 --- SWP Estimated proportion of volunteers asked for
the "How tightly wound spiral arms appear"
question (spiral
-windingproportionvolunteers_asked) (1)
1034-1054 F21.19 --- SAP Estimated proportion of volunteers asked for
the "How many spiral arms" question (spiral
-arm-countproportionvolunteers_asked) (1)
1056-1058 F3.1 --- MP Estimated proportion of volunteers asked for
the "Galaxy merging or disturbed" question
(mergingproportionvolunteers_asked) (1)
1060-1117 A58 --- Fileloc Image file location in the DR1, DR2 or DR5
(file_loc)
1119-1137 F19.15 --- WSS ? Mean distance from center of all pixels
above double the 20th percentile
(wrongsizestatistic)
1139-1143 A5 --- WSW [False True] True if WSS > 161.0 our suggested
starting cut to the mean distance of all
pixels from center (wrongsizewarning)
--------------------------------------------------------------------------------
Note (1): The automated catalogue includes predictions for all galaxies and
all questions even when that question may not be appropriate
(e.g. number of spiral arms for a smooth elliptical). To assess
relevance, we include {question}proportionvolunteers_asked columns
showing the estimated fraction of volunteers that would have been
asked each question (i.e. the product of the vote fractions for the
preceding answers).
We suggest a cut of {question}proportionvolunteers_asked > 0.5 as
a starting point.
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Luc Trabelsi [CDS] 25-Sep-2024