Detector Monitoring with Artificial Neural Networks at the CMS Experiment at the CERN Large Hadron Collider

Abstract

Reliable data quality monitoring is a key asset in delivering collision data suitable for physics analysis in any modern large-scale High Energy Physics experiment. This paper focuses on the use of artificial neural networks for supervised and semi-supervised problems related to the identification of anomalies in the data collected by the CMS muon detectors. We use deep neural networks to analyze LHC collision data, represented as images organized geographically. We train a classifier capable of detecting the known anomalous behaviors with unprecedented efficiency and explore the usage of convolutional autoencoders to extend anomaly detection capabilities to unforeseen failure modes. A generalization of this strategy could pave the way to the automation of the data quality assessment process for present and future high-energy physics experiments.

Searching for new physics with deep autoencoders

“…Moreover there are many other anomaly detection techniques that are not based on autoencoder and/or on reconstruction (loss) which are worth exploring in future work. At the same time autoencoders have been recently used in other high energy physics applications: in parton shower simulation [28], for feature selection of a supervised classification [30], and for automated detection of detector aberrations in CMS [31].…”

“…More recently people have been starting to explore forms of weakly-supervised and unsupervised learning (see e.g. [21][22][23][24][25][26][27][28][29][30][31]). In some weak-supervision approaches, binary classification is attempted on a data sample with only imperfect labels, for instance using class proportions or mixed samples [21][22][23].…”

Variational autoencoders for new physics mining at the Large Hadron Collider

“…In this work, we use VAEs [11,12] based on high-level features as a baseline. Previously, autoencoders have been used in collider physics for detector monitoring [20,21] and event generation [22]. Autoencoders have also been explored to define a jet tagger that would identify new physics events with anomalous jets [23,24], with a strategy similar to what we apply to the full event in this work.…”

Theory Vision at LHCP 2019

“…Another option is to replace theorists with machines, e.g. through weakly supervised or unsupervised learning for classification and anomaly detection [36,37,38,39,40,41,42,43,44,45,46,47]. Although we have some ways to go before these techniques supplant existing search strategies, it is clear that they offer new methods for ensuring that no stone goes unturned at the LHC.…”