Thorsten Glüsenkamp

Thorsten Glüsenkamp

The IceCube Neutrino Observatory is a cubic-kilometer high-energy neutrino detector deployed in the Antarctic ice. Two major event classes are charged-current electron and muon neutrino interactions. In this contribution, we discuss the inference of direction and energy for these classes using conditional normalizing flows. They allow to derive a posterior distribution for each individual event based on the raw data that can include systematic uncertainties, which makes them very promising for next-generation reconstructions. For each normalizing flow we use the differential entropy and the KL-divergence to its maximum entropy approximation to interpret the results. The normalizing flows correctly incorporate complex optical properties of the Antarctic ice and their relation to the embedded detector. For showers, the differential entropy increases in regions of high photon absorption and decreases in clear ice. For muons, the differential entropy strongly correlates with the contained track length. Coverage is maintained, even for low photon counts and highly asymmetrical contour shapes. For high-photon counts, the distributions get narrower and become more symmetrical, as expected from the asymptotic theorem of Bernstein-von-Mises. For shower directional reconstruction, we find the region between 1 TeV and 100 TeV to potentially benefit the most from normalizing flows because of azimuth-zenith asymmetries which have been neglected in previous analyses by assuming symmetrical contours. Events in this energy range play a vital role in the recent discovery of the galactic plane diffuse neutrino emission.

Thorsten Glüsenkamp

The Radio Neutrino Observatory in Greenland (RNO-G) is a radio-based ultra-high energy neutrino detector located at Summit Station, Greenland. It is still being constructed, with 7 stations currently operational. Neutrino detection works by measuring Askaryan radiation produced by neutrino-nucleon interactions. A neutrino candidate must be found amidst other backgrounds which are recorded at much higher rates -- including cosmic-rays and anthropogenic noise -- the origins of which are sometimes unknown. Here we describe a method to classify different noise classes using the latent space of a variational autoencoder. The latent space forms a compact representation that makes classification tractable. We analyze data from a noisy and a silent station. The method automatically detects and allows us to qualitatively separate multiple event classes, including physical wind-induced signals, for both the noisy and the quiet station.

Thorsten Glüsenkamp

A generic D-dimensional Gaussian can be conditioned or projected onto the D-1 unit sphere, thereby leading to the well-known Fisher-Bingham (FB) or Angular Gaussian (AG) distribution families, respectively. These are some of the most fundamental distributions on the sphere, yet cannot straightforwardly be written as a normalizing flow except in two special cases: the von-Mises Fisher in D=3 and the central angular Gaussian in any D. In this paper, we describe how to generalize these special cases to a family of normalizing flows that behave similarly to the full FB or AG family in any D. We call them "zoom-linear-project" (ZLP)-Fisher flows. Unlike a normal Fisher-Bingham distribution, their composition allows to gradually add complexity as needed. Furthermore, they can naturally handle conditional density estimation with target distributions that vary by orders of magnitude in scale - a setting that is important in astronomical applications but that existing flows often struggle with. A particularly useful member of the new family is the Kent analogue that can cheaply upgrade any flow in this situation to yield better performance.

Thorsten Glüsenkamp

Neural-network based predictions of event properties in astro-particle physics are getting more and more common. However, in many cases the result is just utilized as a point prediction. Statistical uncertainties, coverage, systematic uncertainties or a goodness-of-fit measure are often not calculated. Here we describe a certain choice of training and network architecture that allows to incorporate all these properties into a single network model. We show that a KL-divergence objective of the joint distribution of data and labels allows to unify supervised learning and variational autoencoders (VAEs) under one umbrella of stochastic variational inference. The unification motivates an extended supervised learning scheme which allows to calculate a goodness-of-fit p-value for the neural network model. Conditional normalizing flows amortized with a neural network are crucial in this construction. We discuss how to calculate coverage probabilities without numerical integration for specific "base-ordered" contours that are unique to normalizing flows. Furthermore we show how systematic uncertainties can be included via effective marginalization during training. The proposed extended supervised training incorporates (1) coverage calculation, (2) systematics and (3) a goodness-of-fit measure in a single machine-learning model. There are in principle no constraints on the shape of the involved distributions, in fact the machinery works with complex multi-modal distributions defined on product spaces like $\mathbb{R}^n \times \mathbb{S}^m$. The coverage calculation, however, requires care in its interpretation when the distributions are too degenerate. We see great potential for exploiting this per-event information in event selections or for fast astronomical alerts which require uncertainty guarantees.