Unsupervised and lightly supervised learning in particle physics

TL;DR

This paper reviews unsupervised and lightly supervised machine learning techniques in particle physics, highlighting their applications in anomaly detection, detector simulation, and data unfolding, emphasizing their advantages over fully supervised methods.

Contribution

It provides a comprehensive overview of recent unsupervised and semi-supervised machine learning approaches in particle physics, emphasizing their potential and underlying techniques.

Findings

Unsupervised methods excel in anomaly detection by modeling background data.

Generative models significantly speed up detector simulations.

Partially supervised learning aids in data unfolding processes.

Abstract

We review the main applications of machine learning models that are not fully supervised in particle physics, i.e., clustering, anomaly detection, detector simulation, and unfolding. Unsupervised methods are ideal for anomaly detection tasks -- machine learning models can be trained on background data to identify deviations if we model the background data precisely. The learning can also be partially unsupervised when we can provide some information about the anomalies at the data level. Generative models are useful in speeding up detector simulations -- they can mimic the computationally intensive task without large resources. They can also efficiently map detector-level data to parton-level data (i.e., data unfolding). In this review, we focus on interesting ideas and connections and briefly overview the underlying techniques wherever necessary.