Search

Abstract Within the widely used EvtGen framework, we have added a new event generator model for B → K * ℓ + ℓ − with improved standard model (SM) decay amplitudes and possible BSM physics contributions, which are implemented in the operator product expansion in terms of Wilson coefficients. This event generator can then be used to estimate the statistical sensitivity of a simulated experiment to the most general BSM signal resulting from dimension-six operators. We describe the advantages and potential of the newly developed ‘Sibidanov Physics Generator’ in improving the experimental sensitivity of searches for lepton non-universal BSM physics and clarifying signatures. The new generator can properly simulate BSM scenarios, interference between SM and BSM amplitudes, and correlations between different BSM observables as well as acceptance bias. We show that exploiting such correlations substantially improves experimental sensitivity. As a demonstration of the utility of the MC generator, we examine the prospects for improved measurements of lepton non-universality in angular distributions for B → K * ℓ + ℓ − decays from the expected 50 ab −1 data set of the Belle II experiment, using a four-dimensional unbinned maximum likelihood fit. We describe promising experimental signatures and correlations between observables. The use of lepton-universality violating ∆-observables significantly reduces uncertainties in the SM expectations due to QCD and resonance effects and is ideally suited for Belle II with the large data sets expected in the next decade. Thanks to the clean experimental environment of an e + e − machine, Belle II should be able to probe BSM physics in the Wilson coefficients C 7 and C 7 ′ $$ {C}_7^{\prime } $$ , which appear at low q 2 in the di-electron channel.

The status of the research on muon colliders is discussed and plans are outlined for future theoretical and experimental studies. Besides work on the parameters of a 3–4 and 0.5 TeV center-of-mass (COM) energy collider, many studies are now concentrating on a machine near 0.1 TeV (COM) that could be a factory for the s-channel production of Higgs particles. We discuss the research on the various components in such muon colliders, starting from the proton accelerator needed to generate pions from a heavy-Z target and proceeding through the phase rotation and decay (π→μν_{μ}) channel, muon cooling, acceleration, storage in a collider ring, and the collider detector. We also present theoretical and experimental R&D plans for the next several years that should lead to a better understanding of the design and feasibility issues for all of the components. This report is an update of the progress on the research and development since the feasibility study of muon colliders presented at the Snowmass '96 Workshop [R. B. Palmer, A. Sessler, and A. Tollestrup, Proceedings of the 1996 DPF/DPB Summer Study on High-Energy Physics (Stanford Linear Accelerator Center, Menlo Park, CA, 1997)].

We present the first method to probabilistically predict 3D direction in a deep neural network model. The probabilistic predictions are modeled as a heteroscedastic von Mises-Fisher distribution on the sphere $\mathbb{S}^2$ , giving a simple way to quantify aleatoric uncertainty. This approach generalizes the cosine distance loss which is a special case of our loss function when the uncertainty is assumed to be uniform across samples. We develop approximations required to make the likelihood function and gradient calculations stable. The method is applied to the task of predicting the 3D directions of electrons, the most complex signal in a class of experimental particle physics detectors designed to demonstrate the particle nature of dark matter and study solar neutrinos. Using simulated Monte Carlo data, the initial direction of recoiling electrons is inferred from their tortuous trajectories, as captured by the 3D detectors. For $40\,$ keV electrons in a 70% He 30% CO _2 gas mixture at STP, the new approach achieves a mean cosine distance of 0.104 (26 ^∘ ) compared to 0.556 (64 ^∘ ) achieved by a non-machine learning algorithm. We show that the model is well-calibrated and accuracy can be increased further by removing samples with high predicted uncertainty. This advancement in probabilistic 3D directional learning could increase the sensitivity of directional dark matter detectors.