AHAndrew Humphrey

Andrew Humphrey

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Andrew Humphrey has a total of 976 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Euclid Quick Data Release (Q1)

    by B. Margalef-Bentabol, L. Wang, A. La Marca, V. Rodriguez-Gomez, A. Humphrey, S. Fotopoulou, F. Ricci, Y. Toba, G. Stevens, M. Mezcua, et al.

    Active galactic nuclei (AGN) are an important phase in galaxy evolution. However, they can be difficult to identify due to their varied observational signatures. Furthermore, to understand the impact of an AGN on its host galaxy, it is important to quantify the strength of the AGN with respect to the host galaxy. We developed a deep learning (DL) model to identify AGN in imaging data by deriving the contribution of the central point source. The model was trained with Euclidised mock galaxy images in which we artificially injected different levels of AGN, in the form of varying contributions of the point spread function (PSF). Our DL-based method can precisely and accurately recover the injected AGN contribution fraction, f PSF , with a mean difference between the predicted and true f PSF of −0.0078 and an overall root mean square error of 0.051. With this new method, we move beyond the simplistic AGN versus non-AGN classification and are able to precisely quantify the AGN contribution and study galaxy evolution across a continuous spectrum of AGN activity. We applied our method to a stellar-mass-limited sample (with M * ≥ 10 9.8 M ⊙ and 0.5 ≤ z ≤ 2.0 ) selected from the first Euclid quick data release (Q1) and, using a threshold of f PSF > 0.2, we identified 61 432 ± 70 AGN over 63.1 deg 2 (9.8 ± 0.1% of our sample). We compared these DL-selected AGN with AGN selected in the X-ray, mid-infrared (MIR), and via optical spectroscopy and investigated their overlapping fractions depending on different thresholds on the PSF contribution. We find that the overlap increases with increasing X-ray or bolometric AGN luminosity. We observe a positive correlation between the luminosity in the I E filter of the AGN and the host galaxy stellar mass, suggesting that supermassive black holes (SMBHs) generally grow faster (in absolute terms, i.e. the luminosity of the PSF component is larger) in more massive galaxies. Moreover, the mean relative contribution of the AGN is higher in the quiescent galaxy population than in the star-forming population. In terms of absolute power, starburst galaxies, as well as the most massive galaxies (across the star-formation main sequence), tend to host the most luminous AGN, indicating concomitant assembly of the SMBH and the host galaxy.


  • Open Access

    High-resolution radio imaging of TGSS J1530+1049, a radio galaxy in a dense environment at z = 4

    by K. É. Gabányi, S. Frey, L. I. Gurvits, Z. Paragi, K. Perger, A. Saxena, R. A. Overzier, M. Villar-Martín, V. Reynaldi, G. Miley, et al.

    Context. High-redshift radio galaxies can provide important insights into structure formation and galaxy evolution during earlier cosmological epochs. TGSS J1530+1049 was selected as a candidate high-redshift radio galaxy based on its very steep radio spectrum. Subsequent observations with the James Webb Space Telescope ( JWST ) presented in a companion paper have shown that it is located at a redshift of z = 4.0. The JWST data furthermore showed that the radio source is part of one of the densest structures of galaxies and ionized gas known at these redshifts. The complex system qualitatively resembles a massive (cluster) galaxy that formed early through a rapid succession of mergers. Aims. TGSS J1530+1049 is an unresolved source down to an ∼0.6″ scale in multiple radio surveys. To reveal its high-resolution radio structure and allow for a detailed comparison with JWST observations, we studied its morphology at various angular scales with different radio interferometric instruments. Methods. We observed TGSS J1530+1049 at a milliarcsecond- (mas) scale angular resolution with the European VLBI Network (EVN), and at an ∼100-mas scale resolution with the enhanced Multi-Element Remotely Linked Interferometer Network (e-MERLIN). Results. We recovered a complex north–south oriented structure with steep-spectrum radio-emitting features, which are associated with the lobes and hot spots of a jetted active galactic nucleus. However, the centre of the radio galaxy proved to be too faint at centimetre wavelengths to be unambiguously detected in our observations. Nevertheless, its linear size (∼5.5 kpc) and radio power ( L 1.4 GHz ≈ 3 × 10 27 W Hz −1 ) place it among the so-called medium-sized symmetric objects, which are a smaller and/or confined version of larger radio galaxies. A comparison between its radio morphology and that of the ionized gas as observed with the NIRSpec integral field unit on JWST shows that the two are closely aligned. However, the optical emission line gas extends out to ∼25 kpc, which is well beyond the detected radio structures.

  • Open Access

    Euclid Quick Data Release (Q1). Euclid spectroscopy of quasars. 1. Identification and redshift determination of 3500 bright quasars

    by Y. Fu, R. Bouwens, K. I. Caputi, D. Vergani, M. Scialpi, B. Margalef-Bentabol, L. Wang, M. Bolzonella, M. Banerji, E. Bañados, et al.

    The slitless spectroscopy mode of the on board the Euclid telescope has enabled efficient spectroscopy of objects within a large field of view. Nevertheless, the relatively low spectral resolution, overlapping spectra, and contamination pose challenges to source classification and redshift determination using the spectra alone. In this work, we present a large and homogeneous sample of bright quasars identified from the Euclid Quick Data Release (Q1), constructed by combining high-purity candidate selections from to wavelengths without telluric lines, which will be pivotal to quasar spectral analysis. We obtained an empirical spectroscopic depth of JE łesssim 21.5 and ̋E łesssim 21.3 at the sensitivity of the Wide Field Survey, beyond which the number of securely identified quasars declines sharply. Accordingly, the sample presented in this paper comprises spectroscopically confirmed quasars brighter than these limits. We analysed morphological parameters from the Visible Camera (VIS) using Sérsic and model-independent (CAS) metrics, and a deep-learning point spread function fraction to track nuclear dominance. The VIS morphologies show a clear redshift dependence: at low redshift (z<0.5), obvious host structures are common and a single Sérsic model fits about half of the sources; at intermediate redshift (0.5<z<2), the nuclear component dominates, with 90% of the Sérsic fits saturating at the upper index limit. In this intermediate redshift regime, f_ nisp nisp and WISE with the new spectroscopic capabilities of Euclid. Through visual inspection of the Euclid spectra of these quasar candidates, we identify approximately 3500 quasars and determine reliable redshifts in the range of 0< z łesssim4.8. Of these, 2686 are new spectroscopic identifications relative to existing public compilations. We generated the first Euclid composite spectrum of quasars covering rest-frame nuv nir nir PSF is available, and we use it as a more reliable compactness measure than the single-Sérsic and CAS parameters to quantify nuclear versus host emission. We also explore the novel Euclid colour space and discuss the role of these quasars in refining active galactic nucleus selection techniques for future Euclid data releases. nir

  • Open Access

    Anomaly detection in aggregates of grouped time series, composed of high-dimensional and high-frequency event sequences

    by Nuno Miguel Cerqueira da Costa, Armindo Lobo, Suelma Pina, Francisco Franco, Andrew Humphrey, Ali Abbasi, Ricardo Rodrigues, Duarte Fernandes, Negin Mehrbod, Miguel Alves, et al.

    Motivation: As enterprise information systems collect increasingly diverse event streams, the automatic detection of anomalous events is essential for applications like network security, system monitoring, and fraud detection.Methods: This study aims to develop robust techniques for automatically and computationally efficiently identifying anomalous behavior within high-dimensional event sequences, where events occur at specific timestamps (e.g., number of events per hour). The proposed frameworkleverages machine learning to detect anomalies within high-dimensional, high-frequency event sequence time series. By dynamically grouping time series based on shared categorical dimensions, the framework enables scalable anomaly detection across varying levels of aggregation, while maintaining detection sensitivity and specificity.Results: The framework has been validated using real-world telecommunications event data sets and evaluated against the case where no aggregation is performed (‘fully disaggregated’). In the case of our ‘minimum aggregation’ method, the training time is reduced by a factor of∼ 300, and inference time is reduced by a factor of 8. For our other method, ‘ideal aggregation’, the training time is reduced by a factor of 6.5, and the inference time is reduced by a factor of 7.6. The quality of anomaly detection is generally preserved across different aggregation levels of grouped high-dimensional data.Conclusions: This research advances anomaly detection techniques in grouped high-dimensional event sequence data, a complex yet increasingly common data format in fields like network security and system monitoring. The framework’s dimensional aggregation techniques enable efficient pattern recognition while maintaining detection sensitivity and specificity, critical for applications requiring faster anomaly detection response.

  • Open Access

    Euclid preparation

    by A. Humphrey, P. A. C. Cunha, L. Bisigello, C. Tortora, M. Bolzonella, L. Pozzetti, M. Baes, B. R. Granett, A. Amara, S. Andreon, et al.

    The Euclid Space Telescope will image about 14 000 deg2 of the extragalactic sky at visible and near-infrared wavelengths, providing a dataset of unprecedented size and richness that will facilitate a multitude of studies into the evolution of galaxies. Although spectroscopy will also be available for some of the galaxies, in the vast majority of cases the main source of information will come from broadband images and data products thereof (i.e. photometry). Therefore, there is a pressing need to identify or develop scalable yet reliable methodologies to estimate the redshift and physical properties of galaxies using broadband photometry from Euclid. Optionally, such methods could also include ground-based optical photometry. To address this need, we present a novel method developed as part of a ‘data challenge’ within the Euclid Collaboration to estimate the redshift, stellar mass, star-formation rate, specific star-formation rate, E(B − V), and age of galaxies using mock Euclid and ground-based photometry. The main novelty of our property-estimation pipeline is its use of the CatBoost implementation of gradient-boosted regression-trees together with chained regression and an intelligent, automatic optimisation of the training data. The pipeline also includes a computationally efficient method to estimate prediction uncertainties, and, in the absence of ground-truth labels, it provides accurate predictions for metrics of model performance up to z ~ 2. We applied our pipeline to several datasets consisting of mock Euclid broadband photometry and mock ground-based ugriz photometry, with the objective of evaluating the performance of our methodology for estimating the redshift and physical properties of galaxies detected in the Euclid Wide Survey. The statistical metrics of prediction residuals vary depending on which mock catalogue and filters are tested. Nonetheless, the quality of our photometric redshift and physical property estimates are highly competitive overall, validating our modelling approach. However, at z ≳ 3.5, the relative sparsity of galaxies resulted in unreliable redshift and physical property estimates, which we argue could be mitigated by building catalogues with better sampling of z ≳ 3.5 galaxies or by switching to the use of spectral energy distribution fitting in this regime. We also find that the inclusion of ground-based optical photometry significantly improves the quality of the property estimation, highlighting the importance of combining Euclid data with ancillary ground-based data from such surveys as the Vera C. Rubin Observatory Legacy Survey of Space and Time and UNIONS.


  • Open Access

    Euclid preparation. Estimating galaxy physical properties using CatBoost chained regressors with attention

    by Euclid Collaboration, A. Humphrey, P. A. C. Cunha, L. Bisigello, C. Tortora, M. Bolzonella, L. Pozzetti, M. Baes, B. R. Granett, A. Amara, et al.

    Euclid will image ~14000 deg^2 of the extragalactic sky at visible and NIR wavelengths, providing a dataset of unprecedented size and richness that will facilitate a multitude of studies into the evolution of galaxies. In the vast majority of cases the main source of information will come from broad-band images and data products thereof. Therefore, there is a pressing need to identify or develop scalable yet reliable methodologies to estimate the redshift and physical properties of galaxies using broad-band photometry from Euclid, optionally including ground-based optical photometry also. To address this need, we present a novel method to estimate the redshift, stellar mass, star-formation rate, specific star-formation rate, E(B-V), and age of galaxies, using mock Euclid and ground-based photometry. The main novelty of our property-estimation pipeline is its use of the CatBoost implementation of gradient-boosted regression-trees, together with chained regression and an intelligent, automatic optimization of the training data. The pipeline also includes a computationally-efficient method to estimate prediction uncertainties, and, in the absence of ground-truth labels, provides accurate predictions for metrics of model performance up to z~2. We apply our pipeline to several datasets consisting of mock Euclid broad-band photometry and mock ground-based ugriz photometry, to evaluate the performance of our methodology for estimating the redshift and physical properties of galaxies detected in the Euclid Wide Survey. The quality of our photometric redshift and physical property estimates are highly competitive overall, validating our modeling approach. We find that the inclusion of ground-based optical photometry significantly improves the quality of the property estimation, highlighting the importance of combining Euclid data with ancillary ground-based optical data. (Abridged)

  • Open Access

    Exploring the physical properties of type II quasar candidates at intermediate redshifts with CIGALE

    by P. A. C. Cunha, A. Humphrey, J. Brinchmann, A. Paulino-Afonso, L. Bisigello, M. Bolzonella, D. Vaz

    Context. Active galactic nuclei (AGNs) play a vital role in the evolution of galaxies over cosmic time, significantly influencing their star formation and growth. As obscured AGNs are difficult to identify due to obscuration by gas and dust, our understanding of their full impact is still under study. It is essential to investigate their properties and distribution, in particular type II quasars (QSO2s), to comprehensively account for AGN populations and understand how their fraction evolves over time. Such studies provide critical insights into the co-evolution of AGNs and their host galaxies. Aims. Following our previous study, where a machine learning approach was applied to identify 366 QSO2 candidates from SDSS and WISE surveys (median z ∼ 1.1), we now aim to characterise this QSO2 candidate sample by analysing their spectral energy distributions (SEDs) and deriving their physical properties. Methods. We estimated relevant physical properties of the QSO2 candidates, including the star formation rate (SFR), stellar mass (M*), AGN luminosity, and AGN fraction, using SED fitting with CIGALE. We compared the inferred properties with analogous populations in the semi-empirical simulation SPRITZ, placing these results in the context of galaxy evolution. Results. The physical properties derived for our QSO2 candidates indicate a diverse population of AGNs at various stages of evolution. QSO2 candidates cover a wide range in the SFR–M* diagram, with numerous high-SFR sources lying above the main sequence at their redshift, suggesting a link between AGN activity and enhanced star formation. Additionally, we identify a population of apparently quenched galaxies, which may be due to obscured star formation or AGN feedback. Furthermore, the physical parameters of our sample align closely with those of composite systems and type 2 AGNs from SPRITZ, supporting the classification of these candidates as obscured AGNs. Conclusions. This study confirms that our QSO2 candidates, selected via a machine learning approach, exhibit properties consistent with being AGN-host galaxies. This method can identify AGNs within large galaxy samples by considering AGN fractions and their contributions to the infrared luminosity, going beyond the limitations of traditional colour–colour selection techniques. The diverse properties of our candidates demonstrate the capability of this approach to identify complex AGN-host systems that might otherwise be missed. This shows the help that machine learning can provide in refining AGN classifications and advancing our understanding of galaxy evolution driven by AGN activity with new target selection.


  • Open Access

    Euclid Quick Data Release (Q1). Photometric redshifts and physical properties of galaxies through the PHZ processing function

    by Euclid Collaboration, M. Tucci, S. Paltani, W. G. Hartley, F. Dubath, N. Morisset, M. Bolzonella, S. Fotopoulou, F. Tarsitano, C. Saulder, et al.

    The ESA Euclid mission will measure the photometric redshifts of billions of galaxies in order to provide an accurate 3D view of the Universe at optical and near-infrared wavelengths. Photometric redshifts are determined by the PHZ processing function on the basis of the multi-wavelength photometry of Euclid and ground-based observations. In this paper, we describe the PHZ processing used for the Euclid Quick Data Release, the output products, and their validation. The PHZ pipeline is responsible for the following main tasks: source classification into star, galaxy, and QSO classes based on photometric colours; determination of photometric redshifts and of physical properties of galaxies. The classification is able to provide a star sample with a high level of purity, a highly complete galaxy sample, and reliable probabilities of belonging to those classes. The identification of QSOs is more problematic: photometric information seems to be insufficient to accurately separate QSOs from galaxies. The performance of the pipeline in the determination of photometric redshifts has been tested using the COSMOS2020 catalogue and a large sample of spectroscopic redshifts. The results are in line with expectations: the precision of the estimates are compatible with Euclid requirements, while, as expected, a bias correction is needed to achieve the accuracy level required for the cosmological probes. Finally, the pipeline provides reliable estimates of the physical properties of galaxies, in good agreement with findings from the COSMOS2020 catalogue, except for an unrealistically large fraction of very young galaxies with very high specific star-formation rates. The application of appropriate priors is, however, sufficient to obtain reliable physical properties for those problematic objects. We present several areas for improvement for future Euclid data releases.

  • Open Access

    Euclidpreparation

    by A. Humphrey, L. Bisigello, P. A. C. Cunha, M. Bolzonella, S. Fotopoulou, K. Caputi, C. Tortora, G. Zamorani, P. Papaderos, D. Vergani, et al.

    TheEuclidSpace Telescope will provide deep imaging at optical and near-infrared wavelengths, along with slitless near-infrared spectroscopy, across ~15 000deg2of the sky.Euclidis expected to detect ~12 billion astronomical sources, facilitating new insights into cosmology, galaxy evolution, and various other topics. In order to optimally exploit the expected very large dataset, appropriate methods and software tools need to be developed. Here we present a novel machine-learning-based methodology for the selection of quiescent galaxies using broadbandEuclid IE,YE,JE, andHEphotometry, in combination with multi-wavelength photometry from other large surveys (e.g. theRubinLSST). The ARIADNE pipeline uses meta-learning to fuse decision-tree ensembles, nearest-neighbours, and deep-learning methods into a single classifier that yields significantly higher accuracy than any of the individual learning methods separately. The pipeline has been designed to have 'sparsity awareness', such that missing photometry values are informative for the classification. In addition, our pipeline is able to derive photometric redshifts for galaxies selected as quiescent, aided by the 'pseudo-labelling' semi-supervised method, and using an outlier detection algorithm to identify and reject likely catastrophic outliers. After the application of the outlier filter, our pipeline achieves a normalised mean absolute deviation of ≲0.03 and a fraction of catastrophic outliers of ≲0.02 when measured against the COSMOS2015 photometric redshifts. We apply our classification pipeline to mock galaxy photometry catalogues corresponding to three main scenarios: (i)EuclidDeep Survey photometry with ancillaryugriz,WISE, and radio data; (ii)EuclidWide Survey photometry with ancillaryugriz,WISE, and radio data; and (iii)EuclidWide Survey photometry only, with no foreknowledge of galaxy redshifts. In a like-for-like comparison, our classification pipeline outperformsUVJselection, in addition to theEuclid IE–YE,JE–HEandu–IE, IE–JEcolour-colour methods, with improvements in completeness and theF1-score (the harmonic mean of precision and recall) of up to a factor of 2.


  • Improving machine learning-derived photometric redshifts and physical property estimates using unlabelled observations

    by A Humphrey, P A C Cunha, A Paulino-Afonso, S Amarantidis, R Carvajal, J M Gomes, I Matute, P Papaderos

    In the era of huge astronomical surveys, machine learning offers promising solutions for the efficient estimation of galaxy properties. The traditional, ‘supervised’ paradigm for the application of machine learning involves training a model on labelled data, and using this model to predict the labels of previously unlabelled data. The semi-supervised ‘pseudo-labelling’ technique offers an alternative paradigm, allowing the model training algorithm to learn from both labelled data and as-yet unlabelled data. We test the pseudo-labelling method on the problems of estimating redshift, stellar mass, and star formation rate, using COSMOS2015 broad band photometry and one of several publicly available machine learning algorithms, and we obtain significant improvements compared to purely supervised learning. We find that the gradient-boosting tree methods CatBoost, XGBoost, and LightGBM benefit the most, with reductions of up to ∼15 per cent in metrics of absolute error. We also find similar improvements in the photometric redshift catastrophic outlier fraction. We argue that the pseudo-labellng technique will be useful for the estimation of redshift and physical properties of galaxies in upcoming large imaging surveys such as Euclid and LSST, which will provide photometric data for billions of sources.