Related papers: Bayesian Deep-stacking for High-energy Neutrino Se…
The aim of this work is to enable inference of deep networks that retain high accuracy for the least possible model complexity, with the latter deduced from the data during inference. To this end, we revisit deep networks that comprise…
Bayesian neural networks (BNNs) have recently regained a significant amount of attention in the deep learning community due to the development of scalable approximate Bayesian inference techniques. There are several advantages of using a…
Learning Bayesian networks is often cast as an optimization problem, where the computational task is to find a structure that maximizes a statistically motivated score. By and large, existing learning tools address this optimization problem…
Strong-lensing images provide a wealth of information both about the magnified source and about the dark matter distribution in the lens. Precision analyses of these images can be used to constrain the nature of dark matter. However, this…
In this work, we propose a Bayesian thinning algorithm for recovering weighted point source functions in the heat equation from boundary flux observations. The major challenge in the classical Bayesian framework lies in constructing…
A Bayesian approach is presented for detecting and characterising the signal from discrete objects embedded in a diffuse background. The approach centres around the evaluation of the posterior distribution for the parameters of the discrete…
The Bayesian Block algorithm, originally developed for applications in astronomy, can be used to improve the binning of histograms in high energy physics. The visual improvement can be dramatic, as shown here with two simple examples. More…
The IceCube collaboration has reported the first detection of high-energy astrophysical neutrinos including $\sim 50$ high-energy starting events, but no individual sources have been identified. It is therefore important to develop the most…
With the design and development of next-generation high-energy neutrino detectors, it is important to compare different detector designs to optimize detection probability and science reach. These comparisons are nevertheless difficult due…
Ensembles of deep neural networks demonstrate improved performance over single models. For enhancing the diversity of ensemble members while keeping their performance, particle-based inference methods offer a promising approach from a…
Stacking analysis is a means of detecting faint sources using a priori position information to estimate an aggregate signal from individually undetected objects. Confusion severely limits the effectiveness of stacking in deep surveys with…
The sensitivity to dark matter signals at neutrino experiments is fundamentally challenged by the neutrino rates, as they leave similar signatures in their detectors. As a way to improve the signal sensitivity, we investigate a dark matter…
Bayesian neural networks (BNNs) augment deep networks with uncertainty quantification by Bayesian treatment of the network weights. However, such models face the challenge of Bayesian inference in a high-dimensional and usually…
Neutrino astrophysics offers new perspectives on the Universe investigation: high energy neutrinos, produced by the most energetic phenomena in our Galaxy and in the Universe, carry complementary (if not exclusive) information about the…
Neutron-source identification is central to nuclear physics and its applications, from planetary science to nuclear security, yet direct source discrimination from measured neutron spectra remains fundamentally elusive. Here, we introduce a…
Motivation: Modelling methods that find structure in data are necessary with the current large volumes of genomic data, and there have been various efforts to find subsets of genes exhibiting consistent patterns over subsets of treatments.…
This paper proposes a sparse Bayesian treatment of deep neural networks (DNNs) for system identification. Although DNNs show impressive approximation ability in various fields, several challenges still exist for system identification…
A method for a time-dependent search for flaring astrophysical sources which can be potentially detected by large neutrino experiments is presented. The method uses a time-clustering algorithm combined with an unbinned likelihood procedure.…
Bayesian change-point detection, together with latent variable models, allows to perform segmentation over high-dimensional time-series. We assume that change-points lie on a lower-dimensional manifold where we aim to infer subsets of…
I explore the possibility of resurrecting an old, non-Bayesian computational approach for inferring the source direction of a gravitational wave from the output of a two-detector network. The method gives the beam pattern response functions…