Applying Self-organizing Maps to the Inverse Problem

TL;DR

This paper introduces a novel method using self-organizing maps to address inverse problems in particle physics, specifically for identifying hypotheses from observations, and compares its performance with neural networks.

Contribution

It demonstrates that self-organizing maps can effectively compete with neural networks in inverse problem solving without relying on standard model training.

Findings

Self-organizing maps perform competitively against neural networks.

The approach helps characterize observed excesses in particle physics searches.

The method does not require training on standard model processes.

Abstract

In the inverse problem in particle physics, given an unexpected observation, one aims to identify a unique choice from amongst several competing hypotheses. We explore a novel approach of applying self-organizing maps to the inverse problem in a search for vector-like leptons in a trilepton final state. We define an approach combining the inherent clustering of these maps and elements of supervised learning. We compare the performance of this approach with a multiclassfying neural network. We find that the method using self-organizing maps competes well (despite not using any standard model processes in the training), and provides additional tools that would help characterize any observed excesses in searches.