Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 56: Data, AI, Computing, Electronics VI

T 56.2: Talk

Wednesday, March 18, 2026, 16:30–16:45, KH 02.014

b-hive: a CMS wide Machine Learning Framework — Niclas Eich, Alexander Jung, Alexander Schmidt, and •Ulrich Willemsen — III. Physikalisches Institut A, RWTH Aachen

b-hive is a state of the art machine learning framework with wide adaptation in various working groups in CMS. It is a pipeline for training and testing of machine learning algorithms allowing for a modular approach to developing and deploying ML models in high-energy physics applications. The framework provides standardized interfaces for data preprocessing, model training, validation, and inference, enabling researchers to efficiently prototype and compare different algorithms for jet identification and analysis tasks. This presentation will demonstrate the current capabilities of b-hive, showcase recent applications in b-tagging and discuss future development plans.

Keywords: Machine Learning framework; CMS; b-hive