Prof Subir Sarkar

Abstract:

We present GollumFit, a framework designed for performing binned-likelihood analyses on neutrino telescope data. GollumFit incorporates model parameters common to any neutrino telescope and also model parameters specific to the IceCube Neutrino Observatory. We provide a high-level overview of its key features and how the code is organized. We then discuss the performance of the fitting in a typical analysis scenario, highlighting the ability to fit over tens of nuisance parameters. We present some examples showing how to use the package for likelihood minimization tasks. This framework uniquely incorporates the particular model parameters necessary for neutrino telescopes, and solves an associated likelihood problem in a time-efficient manner. PROGRAM Program title: GollumFit Documentation website: https://docs.icecube.aq/gollumfit/main/index.html Developer's repository link: https://github.com/icecube/GollumFit Licensing provisions: GNU Lesser General Public License 2.1 (LGPL) Programming language: C++, Python Nature of problem: Statistical analysis of data from neutrino telescope experiments is often complex and computationally demanding, owing to the need to optimize a likelihood function over many parameters that describe sources of systematic uncertainties and quantities of interest. Solution method: We introduce a framework that performs binned-likelihood optimization, whose performance can handle the number of parameters typical for a neutrino telescope analysis. We highlight a method to perform event-by-event reweighting to incorporate the experimental parameters. In particular, for neutrino telescopes the parameters that incorporate the uncertainties in the atmospheric neutrino flux are common across all experiments and analyses, and are implemented in our framework. The framework has been designed to be easily extendable in the number of observable dimensions and fit parameters. Finally, we use an automatic differentiation package to achieve computational speed in the likelihood optimization.

Abstract:

The IceCube Upgrade is an extension of the existing IceCube Neutrino Observatory and will be deployed in the 2025–2026 austral summer. It will significantly improve the sensitivity of the detector to atmospheric neutrino oscillations. The existing 86-string IceCube array contains a dense in-fill known as DeepCore which is optimized to measure neutrinos with energies down to a few GeV. The IceCube Upgrade will consist of seven new densely instrumented strings placed within the DeepCore volume to further enhance the performance in the GeV energy range. The additional strings will feature new optical modules, each containing multiple photomultiplier tubes (PMTs), in contrast to the existing modules that each contain a single PMT. This will more than triple the number of PMT channels with respect to the current IceCube configuration, allowing for improved detection efficiency and reconstruction performance at GeV energies. We describe necessary updates to simulation, event selection, and reconstruction to accommodate the higher data rates observed by the upgraded detector and the addition of multi-PMT modules. We determine the expected sensitivity of the IceCube Upgrade to the atmospheric neutrino oscillation parameters ${\sin }^2{\theta }_{23}$ and $\mathrm{\Delta }{m}_{32}^2$ , the appearance of tau neutrinos and the neutrino mass ordering. The IceCube Upgrade will provide neutrino oscillation measurements that are of similar precision to those from accelerator experiments, while providing complementarity by probing higher energies and longer baselines, and with different sources of systematic uncertainties.