Search NASA⌕ Search

Engineering topics

Geach, J. E.

Publications and source records attributed to Geach, J. E..

Galaxy morphological classification in deep-wide surveys via unsupervised machine learning

ABSTRACT Galaxy morphology is a fundamental quantity, which is essential not only for the full spectrum of galaxy-evolution studies, but also for a plethora of science in observational cosmology (e.g. as a prior for photometric-redshift measurements and as contextual data for transient light-curve classifications). While a rich literature exists on morphological-classification techniques, the unprecedented data volumes, coupled, in some cases, with the short cadences of forthcoming ‘Big-Data’ surveys (e.g. from the LSST), present novel challenges for this field. Large data volumes make such data sets intractable for visual inspection (even via massively distributed platforms like Galaxy Zoo), while short cadences make it difficult to employ techniques like supervised machine learning, since it may be impractical to repeatedly produce training sets on short time-scales. Unsupervised machine learning, which does not require training sets, is ideally suited to the morphological analysis of new and forthcoming surveys. Here, we employ an algorithm that performs clustering of graph representations, in order to group image patches with similar visual properties and objects constructed from those patches, like galaxies. We implement the algorithm on the Hyper-Suprime-Cam Subaru-Strategic-Program Ultra-Deep survey, to autonomously reduce the galaxy population to a small number (160) of ‘morphological clusters’, populated by galaxies with similar morphologies, which are then benchmarked using visual inspection. The morphological classifications (which we release publicly) exhibit a high level of purity, and reproduce known trends in key galaxy properties as a function of morphological type at z < 1 (e.g. stellar-mass functions, rest-frame colours, and the position of galaxies on the star-formation main sequence). Our study demonstrates the power of unsupervised machine learning in performing accurate morphological analysis, which will become indispensable in this new era of deep-wide surveys.

Martin, G.↗

Concurrent Supermassive Black Hole and Galazy Growth: Linking Environment and Nuclear Activity in Zeta Equals 2.23 H Alpha Emitters

We present results from an approximately equal 100 ks Chandra observation of the 2QZ Cluster 1004+00 structure at z = 2.23 (hereafter 2QZ Clus). 2QZ Clus was originally identified as an overdensity of four optically-selected QSOs at z = 2.23 within a 15 × 15 arcmin square region. Narrow-band imaging in the near-IR (within the K band) revealed that the structure contains an additional overdensity of 22 z = 2.23 H alpha-emitting galaxies (HAEs), resulting in 23 unique z = 2.23 HAEs/QSOs (22 within the Chandra field of view). Our Chandra observations reveal that three HAEs in addition to the four QSOs harbor powerfully accreting supermassive black holes (SMBHs), with 2-10 keV luminosities of approximately equal (8-60) × 10(exp 43) erg s(exp−1) and X-ray spectral slopes consistent with unobscured active galactic nucleus (AGN). Using a large comparison sample of 210 z = 2.23 HAEs in the Chandra-COSMOS field (C-COSMOS), we find suggestive evidence that the AGN fraction increases with local HAE galaxy density. The 2QZ Clus HAEs reside in a moderately overdense environment (a factor of approximately equal 2 times over the field), and after excluding optically-selected QSOs, we find that the AGN fraction is a factor of approximately equal 3.5(+3.8/ −2.2) times higher than C-COSMOS HAEs in similar environments. Using stacking analyses of the Chandra data and Herschel SPIRE observations at 250micrometers, we respectively estimate mean SMBH accretion rates ( M(BH)) and star formation rates (SFRs) for the 2QZ Clus and C-COSMOS samples. We find that the mean 2QZ Clus HAE stacked X-ray luminosity is QSO-like (L(2−10 keV) approximately equal [6-10] × 10(exp 43) erg s(exp −1)), and the implied M(BH)/SFR approximately equal (1.6-3.2) × 10(exp −3) is broadly consistent with the local M(BH)/Stellar Mass relation and z approximately equal 2 X-ray selected AGN. In contrast, the C-COSMOS HAEs are on average an order of magnitude less X-ray luminous and have M(BH)/SFR approximately equal (0.2-0.4) × 10(exp −3), somewhat lower than the local MBH/M relation, but comparable to that found for z approximately equal 1-2 star-forming galaxies with similar mean X-ray luminosities. We estimate that a periodic QSO phase with duty cycle approximately 2%-8% would be sufficient to bring star-forming galaxies onto the local M(BH)/Stellar Mass relation. This duty cycle is broadly consistent with the observed C-COSMOS HAE AGN fraction (Approximately equal 0.4%-2.3%) for powerful AGN with LX approximately greater than 10(exp 44) erg s(exp −1). Future observations of 2QZ Clus will be needed to identify key factors responsible for driving the mutual growth of the SMBHs and galaxies.

Lehmer, B. D.↗