Browsing by Author "Krogager, Jens-Kristian"

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Galaxy Spectra neural Network (GaSNet). II. Using deep learning for spectral classification and redshift predictions
    (2024) Zhong, Fucheng; Napolitano, Nicola R.; Heneka, Caroline; Li, Rui; Bauer, Franz Erik; Bouche, Nicolas; Comparat, Johan; Kim, Young-Lo; Krogager, Jens-Kristian; Longhetti, Marcella; Loveday, Jonathan; Roukema, Boudewijn F.; Rouse, Benedict L.; Salvato, Mara; Tortora, Crescenzo; Assef, Roberto J.; Cassara, Letizia P.; Costantin, Luca; Croom, Scott M.; Davies, Luke J. M.; Fritz, Alexander; Guiglion, Guillaume; Humphrey, Andrew; Pompei, Emanuela; Ricci, Claudio; Sifon, Cristobal; Tempel, Elmo; Zafar, Tayyaba
    The size and complexity reached by the large sky spectroscopic surveys require efficient, accurate, and flexible automated tools for data analysis and science exploitation. We present the Galaxy Spectra Network/GaSNet-II, a supervised multinetwork deep learning tool for spectra classification and redshift prediction. GaSNet-II can be trained to identify a customized number of classes and optimize the redshift predictions. Redshift errors are determined via an ensemble/pseudo-Monte Carlo test obtained by randomizing the weights of the network-of-networks structure. As a demonstration of the capability of GaSNet-II, we use 260k Sloan Digital Sky Survey spectra from Data Release 16, separated into 13 classes including 140k galactic, and 120k extragalactic objects. GaSNet-II achieves 92.4 per cent average classification accuracy over the 13 classes and mean redshift errors of approximately 0.23 per cent for galaxies and 2.1 per cent for quasars. We further train/test the pipeline on a sample of 200k 4MOST (4-metre Multi-Object Spectroscopic Telescope) mock spectra and 21k publicly released DESI (Dark Energy Spectroscopic Instrument) spectra. On 4MOST mock data, we reach 93.4 per cent accuracy in 10-class classification and mean redshift error of 0.55 per cent for galaxies and 0.3 per cent for active galactic nuclei. On DESI data, we reach 96 per cent accuracy in (star/galaxy/quasar only) classification and mean redshift error of 2.8 per cent for galaxies and 4.8 per cent for quasars, despite the small sample size available. GaSNet-II can process similar to 40k spectra in less than one minute, on a normal Desktop GPU. This makes the pipeline particularly suitable for real-time analyses and feedback loops for optimization of Stage-IV survey observations.
  • Thumbnail Image
    Item
    Transverse clues on the kiloparsec-scale structure of the circumgalactic medium as traced by C IV absorption
    (2024) López, Sebastián; Afruni, A.; Zamora Hidalgo, Diego Amaro; Tejos, Nicolás; Ledoux, Cédric; Hernández Guajardo, Joaquín Alexis; Berg, Trystyn A. M.; Cortés Muñoz, Hugo R.; Urbina Jiménez, Francisco; Johnston, E. J.; Barrientos, Luis Felipe; Bayliss, M. B.; Cuellar, Rodrigo; Krogager, Jens-Kristian; Noterdaeme, Pasquier; Solimano, Manuel
    The kiloparsec-scale kinematics and density structure of the circumgalactic medium (CGM) is still poorly constrained observationally, which poses a problem for understanding the role of the baryon cycle in galaxy evolution. Here we present VLT/MUSE integral-field spectroscopy (R ≈ 1800) of four giant gravitational arcs exhibiting W0 ≳ 0.2 Å C IV absorption at eight intervening redshifts, zabs ≈ 2.0–2.5. We detected C IV absorption in a total of 222 adjacent and seeing-uncorrelated sight lines whose spectra sample beams of ("de-lensed") linear size ≈1 kpc. Our data show that (1) absorption velocities cluster at all probed transverse scales, Δr⊥ ≈ 0–15 kpc, depending on system; (2) the (transverse) velocity dispersion never exceeds the mean (line-of-sight) absorption spread; and (3) the (transverse) velocity autocorrelation function does not resolve kinematic patterns at the above spatial scales, but its velocity projection, ξarc(Δv), exhibits a similar shape to the known two-point correlation function toward quasars, ξQSO(Δv). An empirical kinematic model suggests that these results are a natural consequence of wide-beam observations of an unresolved clumpy medium. Our model recovers both the underlying velocity dispersion of the clumps (70–170 km s‑1) and the mean number of clumps per unit area (2–13 kpc‑2). The latter constrains the projected mean inter-clump distance to within ≈0.3–0.8 kpc, which we argue is a measure of clump size for a near-unity covering fraction. The model is also able to predict ξarc(Δv) from ξQSO(Δv), suggesting that the strong systems that shape ξarc(Δv) and the line-of-sight velocity components that define ξQSO(Δv) trace the same kinematic population. Consequently, the clumps must possess an internal density structure that generates both weak and strong components. We discuss how our interpretation is consistent with previous observations using background galaxies and multiple quasars as well as its implications for the connection between the small-scale kinematic structure of the CGM and galactic-scale accretion and feedback processes....