Kristian Zarb Adami
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kristian Zarb Adami has a total of 262 co-authors. Here are the top co-authors by number of shared publications:
- Jackson Levi Said
Institute of Space Sciences and Astronomy, University of Malta, Msida, Malta
10 shared publications - Joseph Sultana
Department of Mathematics, Faculty of Science, University of Malta, MSD2080 Msida, Malta
10 shared publications - Alessio Magro
Institute of Space Sciences and Astronomy, University of Malta, Msida, Malta
6 shared publications - Jurgen Mifsud
Institute of Space Sciences and Astronomy University of Malta Msida Malta
4 shared publications Jackson Levi Said
Physics Department University of Malta 2080 Msida MSD Malta
3 shared publications- Andrea DeMarco
University of Malta
2 shared publications
Latest Publications
- 5 September 2026
- Published by World Scientific Pub Co Pte Ltd in Journal of Astronomical Instrumentation
On the Minimization of Beam Chromaticity in a TEM Horn Antenna for Global 21-cm Cosmology
The detection of the sky-averaged 21-cm signal from the Cosmic Dawn requires antennas whose beam pattern varies minimally with frequency, since beam chromaticity couples the bright, spatially structured Galactic foreground into spurious spectral structure that can obscure the faint cosmological signal. We present a Particle Swarm Optimization (PSO) framework that optimizes a double-ridged transverse-electromagnetic mode (TEM) horn antenna operating from 60 to 85 MHz against a cost function targeting beam chromaticity directly, and benchmark it against both an unoptimized baseline and an established directivity-based cost function [Restrepo, O. A., Lucero, F. I., Chaparro, G. et al. [2023] J. Astron. Instrum. 12, 2350005]. The designs are evaluated by convolving CST-simulated beam patterns with the Global Sky Model to produce antenna-temperature spectra over Local Sidereal Time, and assessed through a mock signal-injection test and a spectral-index analysis. The proposed cost function reduces the directivity smoothness metric by 97% and the beam-solid-angle variation by 52%, and yields the lowest beam-induced foreground residual of the three designs at every foreground order, falling below the depth of an injected 500[Formula: see text]mK signal already at fourth order. In the cold-sky window used for a global-signal integration it is also the least chromatic ([Formula: see text]). We show that the single-power-law spectral index is an incomplete chromaticity metric — its full-day scatter is dominated by Galactic-transit hours excluded from the science integration — and that the foreground-fit residual is the more direct measure of recoverability. These results demonstrate that targeting beam chromaticity directly produces a TEM horn better suited to foreground subtraction than either the baseline or a directivity-based optimization, and provide an observation-driven framework for antenna design in global 21-cm experiments.
- 5 September 2026
- Published by World Scientific Pub Co Pte Ltd in Journal of Astronomical Instrumentation
On the Minimization of Beam Chromaticity in a TEM Horn Antenna for Global 21-cm Cosmology
The detection of the sky-averaged 21-cm signal from the Cosmic Dawn requires antennas whose beam pattern varies minimally with frequency, since beam chromaticity couples the bright, spatially structured Galactic foreground into spurious spectral structure that can obscure the faint cosmological signal. We present a Particle Swarm Optimization (PSO) framework that optimizes a double-ridged transverse-electromagnetic mode (TEM) horn antenna operating from 60 to 85 MHz against a cost function targeting beam chromaticity directly, and benchmark it against both an unoptimized baseline and an established directivity-based cost function [Restrepo, O. A., Lucero, F. I., Chaparro, G. et al. [2023] J. Astron. Instrum. 12, 2350005]. The designs are evaluated by convolving CST-simulated beam patterns with the Global Sky Model to produce antenna-temperature spectra over Local Sidereal Time, and assessed through a mock signal-injection test and a spectral-index analysis. The proposed cost function reduces the directivity smoothness metric by 97% and the beam-solid-angle variation by 52%, and yields the lowest beam-induced foreground residual of the three designs at every foreground order, falling below the depth of an injected 500[Formula: see text]mK signal already at fourth order. In the cold-sky window used for a global-signal integration it is also the least chromatic ([Formula: see text]). We show that the single-power-law spectral index is an incomplete chromaticity metric — its full-day scatter is dominated by Galactic-transit hours excluded from the science integration — and that the foreground-fit residual is the more direct measure of recoverability. These results demonstrate that targeting beam chromaticity directly produces a TEM horn better suited to foreground subtraction than either the baseline or a directivity-based optimization, and provide an observation-driven framework for antenna design in global 21-cm experiments.
- Open Access
- 30 May 2026
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society
- Vol. 549 (3)
Transformer-based source detection and morphological classification in LOFAR Deep-Field continuum images
Radio source detection and morphological classification are fundamental for exploiting the scientific potential of modern radio continuum surveys. However, the rapidly increasing data volumes and the wide diversity of radio morphologies make traditional visual inspection infeasible and pose significant challenges for automated source finding. We apply a transformer-based set-prediction detector (RF-DETR) to 150 MHz continuum images from the LOFAR Deep Fields for instance-level source detection and morphological classification. The method is adapted to multi-frequency-synthesis images of interferometric data and trained with a morphology-driven scheme using five mutually exclusive classes. The model is trained on the ELAIS-N1 Deep Field, where it achieves high detection and classification performance ($\mathrm{F1}\simeq 91$ per cent), and is then applied without retraining to the other three LOFAR Deep Fields. Across all four fields, the model yields consistent catalogues with modest field-to-field differences arising from survey depth and calibration. Compared with widely used PyBDSF catalogues, RF-DETR recovers the majority of PyBDSF sources while representing classical multi-component radio galaxies as single source-level detections rather than fragmented Gaussian components. Artefact-affected and spurious detections are identified as explicit classes, allowing these detections to be distinguished from general astrophysical sources in the resulting catalogues. As external validation, RF-DETR recovers the majority of visually identified extended and giant radio galaxies in the LOFAR Deep Fields and assigns them predominantly to extended morphological classes. These results indicate that transformer-based detectors provide a practical, scalable, morphology-aware approach to source finding in deep radio surveys, with clear relevance for forthcoming facilities such as SKA-Low.
- 30 May 2026
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society
- Vol. 549 (3)
Open AccessTransformer-based source detection and morphological classification in LOFAR Deep-Field continuum images
Radio source detection and morphological classification are fundamental for exploiting the scientific potential of modern radio continuum surveys. However, the rapidly increasing data volumes and the wide diversity of radio morphologies make traditional visual inspection infeasible and pose significant challenges for automated source finding. We apply a transformer-based set-prediction detector (RF-DETR) to 150 MHz continuum images from the LOFAR Deep Fields for instance-level source detection and morphological classification. The method is adapted to multi-frequency-synthesis images of interferometric data and trained with a morphology-driven scheme using five mutually exclusive classes. The model is trained on the ELAIS-N1 Deep Field, where it achieves high detection and classification performance ($\mathrm{F1}\simeq 91$ per cent), and is then applied without retraining to the other three LOFAR Deep Fields. Across all four fields, the model yields consistent catalogues with modest field-to-field differences arising from survey depth and calibration. Compared with widely used PyBDSF catalogues, RF-DETR recovers the majority of PyBDSF sources while representing classical multi-component radio galaxies as single source-level detections rather than fragmented Gaussian components. Artefact-affected and spurious detections are identified as explicit classes, allowing these detections to be distinguished from general astrophysical sources in the resulting catalogues. As external validation, RF-DETR recovers the majority of visually identified extended and giant radio galaxies in the LOFAR Deep Fields and assigns them predominantly to extended morphological classes. These results indicate that transformer-based detectors provide a practical, scalable, morphology-aware approach to source finding in deep radio surveys, with clear relevance for forthcoming facilities such as SKA-Low.
- 15 October 2023
- Published by Springer Nature
- p. 111-114
Planetary Markers in Stellar Spectra: Jupiter-Host Star Classification
- 15 October 2023
- Published by Springer Nature
- p. 111-114
Planetary Markers in Stellar Spectra: Jupiter-Host Star Classification
- Open Access
Comparison of Microwave Hyperthermia Applicator Designs with Fora Dipole and Connected Array
In microwave hyperthermia tumor therapy, electromagnetic waves focus energy on the tumor to elevate the temperature above its normal levels with minimal injury to the surrounding healthy tissue. Microwave hyperthermia applicator design is important for the effectiveness of the therapy and the feasibility of real-time application. In this study, the potential of using fractal octagonal ring antenna elements as a dipole antenna array and as a connected array at 2.45 GHz for breast tumor hyperthermia application was investigated. Microwave hyperthermia treatment models consisting of different fractal octagonal ring antenna array designs and a breast phantom are simulated in COMSOL Multiphysics to obtain the field distributions. The antenna excitation phases and magnitudes are optimized using the global particle swarm algorithm to selectively increase the specific absorption rate at the target region while minimizing hot spots in other regions within the breast. Specific absorption rate distributions, obtained inside the phantom, are analyzed for each proposed microwave hyperthermia applicator design. The dipole fractal octagonal ring antenna arrays are comparatively assessed for three different designs: circular, linear, and Cross—array. The 16-antenna dipole array performance was superior for all three 1-layer applicator designs, and no distinct difference was found between 16-antenna circular, linear, or cross arrays. Two-layer dipole arrays have better performance in the deep-tissue targets than one-layer arrays. The performance of the connected array with a higher number of layers exceeds the performance of the dipole arrays in the superficial regions, while they are comparable for deep regions of the breast. The 1-layer 12-antenna circular FORA dipole array feasibility as a microwave hyperthermia applicator was experimentally shown.
Open AccessComparison of Microwave Hyperthermia Applicator Designs with Fora Dipole and Connected Array
In microwave hyperthermia tumor therapy, electromagnetic waves focus energy on the tumor to elevate the temperature above its normal levels with minimal injury to the surrounding healthy tissue. Microwave hyperthermia applicator design is important for the effectiveness of the therapy and the feasibility of real-time application. In this study, the potential of using fractal octagonal ring antenna elements as a dipole antenna array and as a connected array at 2.45 GHz for breast tumor hyperthermia application was investigated. Microwave hyperthermia treatment models consisting of different fractal octagonal ring antenna array designs and a breast phantom are simulated in COMSOL Multiphysics to obtain the field distributions. The antenna excitation phases and magnitudes are optimized using the global particle swarm algorithm to selectively increase the specific absorption rate at the target region while minimizing hot spots in other regions within the breast. Specific absorption rate distributions, obtained inside the phantom, are analyzed for each proposed microwave hyperthermia applicator design. The dipole fractal octagonal ring antenna arrays are comparatively assessed for three different designs: circular, linear, and Cross—array. The 16-antenna dipole array performance was superior for all three 1-layer applicator designs, and no distinct difference was found between 16-antenna circular, linear, or cross arrays. Two-layer dipole arrays have better performance in the deep-tissue targets than one-layer arrays. The performance of the connected array with a higher number of layers exceeds the performance of the dipole arrays in the superficial regions, while they are comparable for deep regions of the breast. The 1-layer 12-antenna circular FORA dipole array feasibility as a microwave hyperthermia applicator was experimentally shown.
- Open Access
Neural Network Reconstruction of Late-Time Cosmology and Null Tests
The prospect of nonparametric reconstructions of cosmological parameters from observational data sets has been a popular topic in the literature for a number of years. This has mainly taken the form of a technique based on Gaussian processes but this approach is exposed to several foundational issues ranging from overfitting to kernel consistency problems. In this work, we explore the possibility of using artificial neural networks (ANN) to reconstruct late-time expansion and large scale structure cosmological parameters. We first show how mock data can be used to design an optimal ANN for both parameters, which we then use with real data to infer their respective redshift profiles. We further consider cosmological null tests with the reconstructed data in order to confirm the validity of the concordance model of cosmology, in which we observe a mild deviation with cosmic growth data.
Open AccessNeural Network Reconstruction of Late-Time Cosmology and Null Tests
The prospect of nonparametric reconstructions of cosmological parameters from observational data sets has been a popular topic in the literature for a number of years. This has mainly taken the form of a technique based on Gaussian processes but this approach is exposed to several foundational issues ranging from overfitting to kernel consistency problems. In this work, we explore the possibility of using artificial neural networks (ANN) to reconstruct late-time expansion and large scale structure cosmological parameters. We first show how mock data can be used to design an optimal ANN for both parameters, which we then use with real data to infer their respective redshift profiles. We further consider cosmological null tests with the reconstructed data in order to confirm the validity of the concordance model of cosmology, in which we observe a mild deviation with cosmic growth data.
- Open Access
- 28 October 2021
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 655, A5
The Aperture Array Verification System 1: System overview and early commissioning results
The design and development process for the Square Kilometre Array (SKA) radio telescope’s Low Frequency Aperture Array component was progressed during the SKA pre-construction phase by an international consortium, with the goal of meeting requirements for a critical design review. As part of the development process a full-sized prototype SKA Low ‘station’ was deployed – the Aperture Array Verification System 1 (AAVS1). We provide a system overview and describe the commissioning results of AAVS1, which is a low frequency radio telescope with 256 dual-polarisation log-periodic dipole antennas working as a phased array. A detailed system description is provided, including an in-depth overview of relevant sub-systems, ranging from hardware, firmware, software, calibration, and control sub-systems. Early commissioning results cover initial bootstrapping, array calibration, stability testing, beam-forming, and on-sky sensitivity validation. Lessons learned are presented, along with future developments.
- 28 October 2021
- Published by EDP Sciences in Astronomy & Astrophysics
- Vol. 655, A5
Open AccessThe Aperture Array Verification System 1: System overview and early commissioning results
The design and development process for the Square Kilometre Array (SKA) radio telescope’s Low Frequency Aperture Array component was progressed during the SKA pre-construction phase by an international consortium, with the goal of meeting requirements for a critical design review. As part of the development process a full-sized prototype SKA Low ‘station’ was deployed – the Aperture Array Verification System 1 (AAVS1). We provide a system overview and describe the commissioning results of AAVS1, which is a low frequency radio telescope with 256 dual-polarisation log-periodic dipole antennas working as a phased array. A detailed system description is provided, including an in-depth overview of relevant sub-systems, ranging from hardware, firmware, software, calibration, and control sub-systems. Early commissioning results cover initial bootstrapping, array calibration, stability testing, beam-forming, and on-sky sensitivity validation. Lessons learned are presented, along with future developments.
- Open Access
Reconstructing teleparallel gravity with cosmic structure growth and expansion rate data
In this work, we use a combined approach of Hubble parameter data together with redshift-space-distortion $(f\sigma_8)$ data, which together are used to reconstruct the teleparallel gravity (TG) Lagrangian via Gaussian processes (GP). The adopted Hubble data mainly comes from cosmic chronometers, while for the Type Ia supernovae data we use the latest jointly calibrated Pantheon compilation. Moreover, we consider two main GP covariance functions, namely the squared-exponential and Cauchy kernels in order to show consistency (to within 1$\sigma$ uncertainties). The core results of this work are the numerical reconstructions of the TG Lagrangian from GP reconstructed Hubble and growth data. We take different possible combinations of the datasets and kernels to illustrate any potential differences in this regard. We show that nontrivial cosmology beyond $\Lambda$CDM falls within the uncertainties of the reconstructions from growth data, which therefore indicates no significant departure from the concordance cosmological model.Open AccessReconstructing teleparallel gravity with cosmic structure growth and expansion rate data
In this work, we use a combined approach of Hubble parameter data together with redshift-space-distortion $(f\sigma_8)$ data, which together are used to reconstruct the teleparallel gravity (TG) Lagrangian via Gaussian processes (GP). The adopted Hubble data mainly comes from cosmic chronometers, while for the Type Ia supernovae data we use the latest jointly calibrated Pantheon compilation. Moreover, we consider two main GP covariance functions, namely the squared-exponential and Cauchy kernels in order to show consistency (to within 1$\sigma$ uncertainties). The core results of this work are the numerical reconstructions of the TG Lagrangian from GP reconstructed Hubble and growth data. We take different possible combinations of the datasets and kernels to illustrate any potential differences in this regard. We show that nontrivial cosmology beyond $\Lambda$CDM falls within the uncertainties of the reconstructions from growth data, which therefore indicates no significant departure from the concordance cosmological model. - 9 June 2021
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society
- Vol. 505 (4) , 6155-6165
A comparative study of convolutional neural networks for the detection of strong gravitational lensing
As we enter the era of large-scale imaging surveys with the upcoming telescopes such as the Large Synoptic Survey Telescope (LSST) and the Square Kilometre Array (SKA), it is envisaged that the number of known strong gravitational lensing systems will increase dramatically. However, these events are still very rare and require the efficient processing of millions of images. In order to tackle this image processing problem, we present machine learning techniques and apply them to the gravitational lens finding challenge. The convolutional neural networks (CNNs) presented here have been reimplemented within a new, modular, and extendable framework, Lens EXtrActor CaTania University of Malta (LEXACTUM). We report an area under the curve (AUC) of 0.9343 and 0.9870, and an execution time of 0.0061 and 0.0594 s per image, for the Space and Ground data sets, respectively, showing that the results obtained by CNNs are very competitive with conventional methods (such as visual inspection and arc finders) for detecting gravitational lenses.
- 9 June 2021
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society
- Vol. 505 (4) , 6155-6165
A comparative study of convolutional neural networks for the detection of strong gravitational lensing
As we enter the era of large-scale imaging surveys with the upcoming telescopes such as the Large Synoptic Survey Telescope (LSST) and the Square Kilometre Array (SKA), it is envisaged that the number of known strong gravitational lensing systems will increase dramatically. However, these events are still very rare and require the efficient processing of millions of images. In order to tackle this image processing problem, we present machine learning techniques and apply them to the gravitational lens finding challenge. The convolutional neural networks (CNNs) presented here have been reimplemented within a new, modular, and extendable framework, Lens EXtrActor CaTania University of Malta (LEXACTUM). We report an area under the curve (AUC) of 0.9343 and 0.9870, and an execution time of 0.0061 and 0.0594 s per image, for the Space and Ground data sets, respectively, showing that the results obtained by CNNs are very competitive with conventional methods (such as visual inspection and arc finders) for detecting gravitational lenses.
- 1 June 2021
- Published by IOP Publishing in Journal of Cosmology and Astroparticle Physics
- Vol. 2021 (06) , 015
Reconstructing teleparallel gravity with cosmic structure growth and expansion rate data
In this work, we use a combined approach of Hubble parameter data together with redshift-space-distortion (fσ8) data, which together are used to reconstruct the teleparallel gravity (TG) Lagrangian via Gaussian processes (GP). The adopted Hubble data mainly comes from cosmic chronometers, while for the Type Ia supernovae data we use the latest jointly calibrated Pantheon compilation. Moreover, we consider two main GP covariance functions, namely the squared-exponential and Cauchy kernels in order to show consistency (to within 1σ uncertainties). The core results of this work are the numerical reconstructions of the TG Lagrangian from GP reconstructed Hubble and growth data. We take different possible combinations of the datasets and kernels to illustrate any potential differences in this regard. We show that nontrivial cosmology beyond ΛCDM falls within the uncertainties of the reconstructions from growth data, which therefore indicates no significant departure from the concordance cosmological model.
- 1 June 2021
- Published by IOP Publishing in Journal of Cosmology and Astroparticle Physics
- Vol. 2021 (06) , 015
Reconstructing teleparallel gravity with cosmic structure growth and expansion rate data
In this work, we use a combined approach of Hubble parameter data together with redshift-space-distortion (fσ8) data, which together are used to reconstruct the teleparallel gravity (TG) Lagrangian via Gaussian processes (GP). The adopted Hubble data mainly comes from cosmic chronometers, while for the Type Ia supernovae data we use the latest jointly calibrated Pantheon compilation. Moreover, we consider two main GP covariance functions, namely the squared-exponential and Cauchy kernels in order to show consistency (to within 1σ uncertainties). The core results of this work are the numerical reconstructions of the TG Lagrangian from GP reconstructed Hubble and growth data. We take different possible combinations of the datasets and kernels to illustrate any potential differences in this regard. We show that nontrivial cosmology beyond ΛCDM falls within the uncertainties of the reconstructions from growth data, which therefore indicates no significant departure from the concordance cosmological model.
- 23 November 2020
- Published by The Royal Society in Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
- Vol. 379 (2188) , 20190575
Low-frequency technology for a lunar interferometer
This work sketches a possible design architecture of a low-frequency radio interferometer located on the lunar surface. The design has evolved from single antenna experiments aimed at the global signal detection of the epoch of reionization (EoR) to the square kilometre array (SKA) which, when complete, will be capable of imaging the highly red-shifted H 1 -signal from the cosmic dawn through to the EoR. However, due to the opacity of the ionosphere below 10 MHz and the anthropogenic radio-frequency interference, these terrestrial facilities are incapable of detecting pre-ionization signals and the moon becomes an attractive location to build a low-frequency radio interferometer capable of detecting such cosmological signals. Even though there are enormous engineering challenges to overcome, having this scientific facility on the lunar surface also opens up several new exciting possibilities for low-frequency radio astronomy. This article is part of a discussion meeting issue ‘Astronomy from the Moon: the next decades (part 1)’.
- 23 November 2020
- Published by The Royal Society in Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
- Vol. 379 (2188) , 20190575
Low-frequency technology for a lunar interferometer
This work sketches a possible design architecture of a low-frequency radio interferometer located on the lunar surface. The design has evolved from single antenna experiments aimed at the global signal detection of the epoch of reionization (EoR) to the square kilometre array (SKA) which, when complete, will be capable of imaging the highly red-shifted H 1 -signal from the cosmic dawn through to the EoR. However, due to the opacity of the ionosphere below 10 MHz and the anthropogenic radio-frequency interference, these terrestrial facilities are incapable of detecting pre-ionization signals and the moon becomes an attractive location to build a low-frequency radio interferometer capable of detecting such cosmological signals. Even though there are enormous engineering challenges to overcome, having this scientific facility on the lunar surface also opens up several new exciting possibilities for low-frequency radio astronomy. This article is part of a discussion meeting issue ‘Astronomy from the Moon: the next decades (part 1)’.