17–24 Jul 2024
Prague
Europe/Prague timezone

Tracker Data Quality Certification in CMS with new Machine Learning Tools

20 Jul 2024, 17:53
17m
Club A

Club A

Parallel session talk 14. Computing, AI and Data Handling Computing and Data handling

Speaker

Atanu Pathak (Purdue University Northwest (US))

Description

The CMS Tracker in Run 3 is made up of thousands of silicon modules (Pixel:1856 modules, Strip: 15148 modules). Because of the aging of the detector, and all other possible accidents that may happen during the operations, there is the need for constant monitoring of the detector components, in order to guarantee the best data quality. The procedures and tools adopted by the CMS Tracker group to monitor the data (both Online and Offline) will be presented, together with the results of the data certification in Run 3. New tools are being developed to support the work of the human shifters in the daily task of checking hundreds of histograms. To further increase the efficacy of the process, machine learning models are being trained on data to identify the most subtle discrepancies that may pass unnoticed to the human eye.

I read the instructions above Yes

Primary authors

Atanu Pathak (Purdue University Northwest (US)) CMS

Presentation materials