Abstract: Aims. Traditional star-galaxy classification techniques often rely on feature estimation from catalogs, a process susceptible to introducing inaccuracies, thereby potentially jeopardizing the classification's reliability. Certain galaxies, especially those not manifesting as extended sources, can be misclassified when their shape parameters and flux solely drive the inference. We aim to create a robust and accurate classification network for identifying stars and galaxies directly from astronomical images.Methods. The AutoSourceID-Classifier (ASID-C) algorithm developed for this work uses 32x32 pixel single filter band source cutouts generated by the previously developed AutoSourceID-Light (ASID-L) code. By leveraging convolutional neural networks (CNN) and additional information about the source position within the full-field image, ASID-C aims to accurately classify all stars and galaxies within a survey. Subsequently, we employed a modified Platt scaling calibration for the output of the CNN, ensuring that the derived probabilities were effectively calibrated, delivering precise and reliable results.Results. We show that ASID-C, trained on MeerLICHT telescope images and using the Dark Energy Camera Legacy Survey (DECaLS) morphological classification, is a robust classifier and outperforms similar codes such as SourceExtractor. To facilitate a rigorous comparison, we also trained an eXtreme Gradient Boosting (XGBoost) model on tabular features extracted by SourceExtractor. While this XGBoost model approaches ASID-C in performance metrics, it does not offer the computational efficiency and reduced error propagation inherent in ASID-C's direct image-based classification approach. ASID-C excels in low signal-to-noise ratio and crowded scenarios, potentially aiding in transient host identification and advancing deep-sky astronomy.

Abstract: Cosmological bounds on neutrinos and additional hypothetical light thermal relics, such as QCD axions, are currently among the most restrictive ones. These limits mainly rely on cosmic microwave background temperature anisotropies. Nonetheless, one of the largest cosmological signatures of thermal relics is that on gravitational lensing, due to their free-streaming behavior before their nonrelativistic period. We investigate late-time only hot-relic mass constraints, primarily based on recently released lensing data from the Atacama Cosmology Telescope, both alone and in combination with lensing data from the Planck satellite. Additionally, we consider other local probes, such as baryon acoustic oscillations measurements, shear-shear, galaxy-galaxy, and galaxy-shear correlation functions from the dark energy survey, and distance moduli measurements from Type-Ia Supernovae. The tightest bounds we find are Sigma m(v) < 0.43 eV and m(a) < 1.1 eV, both at 95% CL Interestingly, these limits are still much stronger than those found on e.g., laboratory neutrino mass searches, reassessing the robustness of the extraction of thermal relic properties via cosmological observations. In addition, when considering lensing-only data, the significance of the Hubble constant tension is considerably reduced, while the clustering parameter sigma 8 controversy is completely absent.

Abstract: A search for a new heavy scalar particle X decaying into a Standard Model (SM) Higgs boson and a new singlet scalar particle S is presented. The search uses a proton-proton (pp) collision data sample with an integrated luminosity of 140 fb(-1) recorded at a centre-of-mass energy of root s = 13 TeV with the ATLAS detector at the Large Hadron Collider. The most sensitive mass parameter space is explored in X mass ranging from 500 to 1500 GeV, with the corresponding S mass in the range 200-500 GeV. The search selects events with two hadronically decaying tau-lepton candidates from H -> tau(+)tau(-) decays and one or two light leptons (l = e, mu) from S -> VV (V = W, Z) decays while the remaining V boson decays hadronically or to neutrinos. A multivariate discriminant based on event kinematics is used to separate the signal from the background. No excess is observed beyond the expected SM background and 95% confidence level upper limits between 72 fb and 542 fb are derived on the cross-section sigma(pp -> X -> SH) assuming the same SM-Higgs boson-like decay branching ratios for the S -> VV decay. Upper limits on the visible cross-sections sigma(pp -> X -> SH -> WW tau tau) and sigma(pp -> X -> SH -> ZZ tau tau) are also set in the ranges 3-26 fb and 6-33 fb, respectively.