Abstract: Context. At GeV energies, the sky is dominated by the interstellar emission from the Galaxy. With limited statistics and spatial resolution, accurately separating point sources is therefore challenging. Aims. Here we present the first application of deep learning based algorithms to automatically detect and classify point sources from gamma-ray data. For concreteness we refer to this approach as AutoSourceID. Methods. To detect point sources, we utilized U-shaped convolutional networks for image segmentation and k-means for source clustering and localization. We also explored the Centroid-Net algorithm, which is designed to find and count objects. Using two algorithms allows for a cross check of the results, while a combination of their results can be used to improve performance. The training data are based on 9.5 years of exposure from The Fermi Large Area Telescope (Fermi-LAT) and we used source properties of active galactic nuclei (AGNs) and pulsars (PSRs) from the fourth Fermi-LAT source catalog in addition to several models of background interstellar emission. The results of the localization algorithm are fed into a classification neural network that is trained to separate the three general source classes (AGNs, PSRs, and FAKE sources). Results. We compared our localization algorithms qualitatively with traditional methods and find them to have similar detection thresholds. We also demonstrate the robustness of our source localization algorithms to modifications in the interstellar emission models, which presents a clear advantage over traditional methods. The classification network is able to discriminate between the three classes with typical accuracy of similar to 70%, as long as balanced data sets are used in classification training. We published online our training data sets and analysis scripts and invite the community to join the data challenge aimed to improve the localization and classification of gamma-ray point sources.

Abstract: Aims. In astronomy, machine learning has been successful in various tasks such as source localisation, classification, anomaly detection, and segmentation. However, feature regression remains an area with room for improvement. We aim to design a network that can accurately estimate sources' features and their uncertainties from single-band image cutouts, given the approximated locations of the sources provided by the previously developed code AutoSourceID-Light (ASID-L) or other external catalogues. This work serves as a proof of concept, showing the potential of machine learning in estimating astronomical features when trained on meticulously crafted synthetic images and subsequently applied to real astronomical data.Methods. The algorithm presented here, AutoSourceID-FeatureExtractor (ASID-FE), uses single-band cutouts of 32x32 pixels around the localised sources to estimate flux, sub-pixel centre coordinates, and their uncertainties. ASID-FE employs a two-step mean variance estimation (TS-MVE) approach to first estimate the features and then their uncertainties without the need for additional information, for example the point spread function (PSF). For this proof of concept, we generated a synthetic dataset comprising only point sources directly derived from real images, ensuring a controlled yet authentic testing environment.Results. We show that ASID-FE, trained on synthetic images derived from the MeerLICHT telescope, can predict more accurate features with respect to similar codes such as SourceExtractor and that the two-step method can estimate well-calibrated uncertainties that are better behaved compared to similar methods that use deep ensembles of simple MVE networks. Finally, we evaluate the model on real images from the MeerLICHT telescope and the Zwicky Transient Facility (ZTF) to test its transfer learning abilities.

Abstract: We present an efficient method of reconstructing the parameters of the constrained MSSM from assumed future LHC data, applied both on their own right and in combination with the cosmological determination of the relic dark matter abundance. Focusing on the ATLAS SU3 benchmark point, we demonstrate that our simple Gaussian approximation can recover the values of its parameters remarkably well. We examine two popular noninformative priors and obtain very similar results, although when we use an informative, naturalness-motivated prior, we find some sizeable differences. We show that a further strong improvement in reconstructing the SU3 parameters can by achieved by applying additional information about the relic abundance at the level of WMAP accuracy, although the expected data from Planck will have only a very limited additional impact. Further external data may be required to break some remaining degeneracies. We argue that the method presented here is applicable to a wide class of low-energy effective supersymmetric models, as it does not require one to deal with purely experimental issues, e.g., detector performance, and has the additional advantages of computational efficiency. Furthermore, our approach allows one to distinguish the effect of the model's internal structure and of the external data on the final parameters constraints.

Abstract: We carry out global fits to the nonuniversal Higgs Model (NUHM), applying all relevant present-day constraints. We present global probability maps for the NUHM parameters and observables (including collider signatures, direct, and indirect detection quantities), both in terms of posterior probabilities and in terms of profile likelihood maps. We identify regions of the parameter space where the neutralino dark matter in the model is either binolike, or else higgsinolike with mass close to 1 TeV and a spin-independent scattering cross section similar to 10(-9)-10(-8) pb. We trace the occurrence of the higgsinolike region to be a consequence of a mild focusing effect in the running of one of the Higgs masses, the existence of which in the NUHM we identify in our analysis. Although the usual binolike neutralino is more prominent, higgsinolike dark matter cannot be excluded, however its significance strongly depends on the prior and statistics used to assess it. We note that, despite experimental constraints often favoring different regions of parameter space to the constrained minimal supersymmetric standard model, most observational consequences appear fairly similar, which will make it challenging to distinguish the two models experimentally.

Abstract: We investigate the reconstruction capabilities of the dark matter mass and spin-independent cross section from future ton-scale direct detection experiments using germanium, xenon, or argon as targets. Adopting realistic values for the exposure, energy threshold, and resolution of dark matter experiments which will come online within 5 to 10 years, the degree of complementarity between different targets is quantified. We investigate how the uncertainty in the astrophysical parameters controlling the local dark matter density and velocity distribution affects the reconstruction. For a 50 GeV WIMP, astrophysical uncertainties degrade the accuracy in the mass reconstruction by up to a factor of similar to 4 for xenon and germanium, compared to the case when astrophysical quantities are fixed. However, the combination of argon, germanium, and xenon data increases the constraining power by a factor of similar to 2 compared to germanium or xenon alone. We show that future direct detection experiments can achieve self-calibration of some astrophysical parameters, and they will be able to constrain the WIMP mass with only very weak external astrophysical constraints.