Binary junipr: An Interpretable Probabilistic Model for Discrimination.

Anders Andreassen,Matthew D Schwartz,Christopher Frye,Ilya Feige

doi:10.1103/physrevlett.123.182001

Abstract

junipr is an approach to unsupervised learning in particle physics that scaffolds a probabilistic model for jets around their representation as binary trees. Separate junipr models can be learned for different event or jet types, then compared and explored for physical insight. The relative probabilities can also be used for discrimination. In this Letter, we show how the training of the separate models can be refined in the context of classification to optimize discrimination power. We refer to this refined approach as binary junipr. binary junipr achieves state-of-the-art performance for quark-gluon discrimination and top tagging. The trained models can then be analyzed to provide physical insight into how the classification is achieved. As examples, we explore differences between quark and gluon jets and between gluon jets generated with two different simulations.

Highlights

Modern machine learning has already made impressive contributions to particle physics
A key question that is beginning to be addressed is this: what is the optimal representation of the information in an event? Is it through analogy with images [1,2], natural-language processing [8,11], or set theory [19,20]? In many of these approaches, there is a competition between effectiveness in some task and interpretability of the neural network
An approach to machine learning for particle physics called JUNIPR [21] builds a separate network for each jet type using a physical representation of the information in the jet: the jet clustering tree

Summary

Introduction

Modern machine learning has already made impressive contributions to particle physics. An approach to machine learning for particle physics called JUNIPR [21] builds a separate network for each jet type using a physical representation of the information in the jet: the jet clustering tree. To learn these probability distributions, JUNIPR introduces a quantity hðtÞ as a representation of fkð1tÞ; ...; kðttÞg, i.e., the “state” of the jet at branching step t.

Results

Conclusion