AutoDQM Tool for CMS Data Certification

The Compact Muon Solenoid (CMS) detector collects a large amount of data for analysis. To confirm that the collected data is acceptable for physics analysis, each run is certified by CMS data shifters. One requirement of this task is to compare the newly reconstructed data to a high quality run collected previously. This is a very time-consuming and labor-intensive task. The AutoDQM group created a web tool that aims to semi-automate this task as well as to improve the accuracy of the comparison. We utilize various machine learning algorithms and statistical tests to compare a large number of plots simultaneously and flag only plots that are anomalous and require further inspection. During Run 3 data-taking in 2022, trigger shifters began incorporating AutoDQM into their certification workflow. In this presentation, I give an overview of the AutoDQM tool, its workflow, and the machine learning algorithms and statistical tests employed within the tool.