Search

Measurements of Ultra-High Energy Cosmic Rays (UHECR) suggest a complex composition with significant contributions from heavy nuclei at the highest energies. We systematically explore how the selection and number of primary nuclei included in the analysis impact the inferred UHECR mass composition. Introducing a distance measure in the space of $X_{\rm max}$ distribution moments, we demonstrate that limiting the analysis to a few primaries can introduce significant biases, particularly as observational data improves. We provide lists of primaries approximately equidistant in the new measure, which guaranty unbiased results at given statistical confidence. Additionally, we explore consistent inclusion of nuclei heavier than iron and up to plutonium, deriving first observational upper bounds on their contributions to UHECR with the Pierre Auger Open Data., Comment: 11 pages, 8 figures

We show that the maximum shower depth distributions of Ultra-High Energy Cosmic Rays (UHECRs), as measured by fluorescence telescopes, can be augmented by building a mapping to observables collected by surface detectors. The resulting statistical improvement of such augmented dataset depends in a universal way on the strength of the correlation exhibited by the mapping. Building upon the publicly available data on "golden hybrid" events from the Pierre Auger Observatory we project possible improvements in the inferred composition of UHECRs for a range of possible mappings with varying correlation strengths., Comment: 8 pages, 6 figures, method corrected

We apply statistical inference on the Pierre Auger Open Data to discern for the first time the full mass composition of cosmic rays at different energies. Working with longitudinal electromagnetic profiles of cosmic ray showers, in particular their peaking depths $X_{\rm max}$, we employ central moments of the $X_{\rm max}$ distributions as features to discriminate between different shower compositions. We find that already the first few moments entail the most relevant information to infer the primary cosmic ray mass spectrum. Our approach, based on an unbinned likelihood, allows us to consistently account for sources of statistical uncertainties due to finite datasets, both measured and simulated, as well as systematic effects. Finally, we provide a quantitative comparison of different high energy hadronic interaction models available in the atmospheric shower simulation codes., Comment: 28 pages + 16 page Appendix, 27 figures

Unsupervised anomaly detection could be crucial in future analyses searching for rare phenomena in large datasets, as for example collected at the LHC. To this end, we introduce a physics inspired variational autoencoder (VAE) architecture which performs competitively and robustly on the LHC Olympics Machine Learning Challenge datasets. We demonstrate how embedding some physical observables directly into the VAE latent space, while at the same time keeping the classifier manifestly agnostic to them, can help to identify and characterise features in measured spectra as caused by the presence of anomalies in a dataset., Comment: Version as published

A new paradigm for data-driven, model-agnostic new physics searches at colliders is emerging, and aims to leverage recent breakthroughs in anomaly detection and machine learning. In order to develop and benchmark new anomaly detection methods within this framework, it is essential to have standard datasets. To this end, we have created the LHC Olympics 2020, a community challenge accompanied by a set of simulated collider events. Participants in these Olympics have developed their methods using an R&D dataset and then tested them on black boxes: datasets with an unknown anomaly (or not). This paper will review the LHC Olympics 2020 challenge, including an overview of the competition, a description of methods deployed in the competition, lessons learned from the experience, and implications for data analyses with future datasets as well as future colliders., Comment: 108 pages, 53 figures, 3 tables

We reconsider decays of pseudoscalar mesons (P) to neutrino pairs and possibly additional photons in presence of (light) Majorana neutrinos. For this purpose we derive a model-independent general parametrization of neutrino mass matrices with physically interpretable and irreducible set of parameters. The parametrization is valid for any number of neutrinos and interpolates smoothly between the heavy Majorana and the (pseudo)Dirac neutrino limits. We apply the new parametrization to the study of $P \rightarrow \nu \nu$ and $P \rightarrow \nu \nu \gamma$ decays within the SM extended by additional singlet fermions. We update the SM predictions for the branching ratios of $B_{s,d} \rightarrow \nu \nu \gamma$ and discuss the sensitivity of the $B_{s,d} \rightarrow E_{\rm miss} (\gamma)$ decays to neutrino mass and mixing parameters., Comment: 20 pages, 6 figures, added one appendix, conclusions unchanged, version as published

We use machine learning (ML) and non-ML techniques to study optimized $CP$-odd observables, directly and maximally sensitive to the $CP$-odd $i \tilde \kappa \bar t \gamma^5 t h$ interaction at the LHC and prospective future hadron colliders using the final state with a Higgs boson and a top quark pair, $pp\to t\bar t h$, followed by semileptonic $t$ decays. We perform phase-space optimization of manifestly $CP$-odd observables ($\boldsymbol \omega$), sensitive to the sign of $\tilde \kappa$, and constructed from experimentally accessible final state momenta. We identify a simple optimized linear combination $\boldsymbol \alpha\cdot \boldsymbol\omega$ that gives similar sensitivity as the studied fully fledged ML models. Using $\boldsymbol\alpha\cdot \boldsymbol\omega$ we project the expected sensitivities to $\tilde \kappa$ at HL-LHC, HE-LHC, and FCC-hh., Comment: As published