Fast Identification of Transients: Applying Expectation Maximization to Neutrino Data

TL;DR

This paper introduces a fast, unsupervised Expectation Maximization-based method for detecting transient neutrino signals, significantly reducing computation time and adaptable to multiple flares, demonstrated on IceCube data.

Contribution

The paper presents a novel EM algorithm tailored for transient detection in neutrino data, offering over 10,000-fold speed improvements and flexibility for multiple flare analysis.

Findings

Achieved over 10,000 times faster computation on a single CPU.

Successfully applied the method to IceCube neutrino flare data.

Demonstrated extension capability to multiple flare analysis.

Abstract

We present a novel method for identifying transients suitable for both strong signal-dominated and background-dominated objects. By employing the unsupervised machine learning algorithm known as Expectation Maximization, we achieve computing time reductions of over $10^4$ on a single CPU compared to conventional brute-force methods. Furthermore, this approach can be readily extended to analyze multiple flares. We illustrate the algorithm's application by fitting the IceCube neutrino flare of TXS 0506+056.