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

T: Fachverband Teilchenphysik

T 28: Data, AI, Computing, Electronics III

T 28.9: Talk

Tuesday, March 17, 2026, 18:15–18:30, KH 00.024

Keeping Track of Graphs: 4D Tracking with Graph Neural Networks at Muon Colliders — •Lukas Bauckhage — Deutsches Elektronen-Synchrotron DESY — Physikalisches Institut, Universität Bonn

This talk explores the application of Graph Neural Networks (GNNs) to track reconstruction at a future muon collider experiment utilising precise timing information. We highlight the challenges posed by high beam-induced backgrounds, making robust and efficient tracking exceptionally difficult. We demonstrate how GNNs can effectively model the 4D relationships among detector hits to identify and group related hits by leveraging spatial and timing information to distinguish true particle trajectories from background activity. Our results on the performance of reconstruction algorithms aided by GNNs compared to established algorithms purely based on Kalman-Filters are presented.

Keywords: GNN; 4D tracking; muon collider; tracking