Geometric Graph Neural Network based track finding

23 Sept 2024 (modified: 05 Feb 2025)Submitted to ICLR 2025EveryoneRevisionsBibTeXCC BY 4.0
Keywords: Tracking, GNN, High Energy Physics
Abstract:

An essential component of event reconstruction in particle physics experiments is identifying the trajectory of charged particles in the detector. Traditional methods for track finding are often complex, and tailored to specific detectors and input geometries, limiting their adaptability to new detector designs and optimization processes. To overcome these limitations, we present a novel, end-to-end track finding algorithm that is detector-agnostic and can take into account multiple input geometric types. To achieve this, our approach unifies inputs from multiple sub-detectors and detector types into a single geometric algebra representation, simplifying data handling compared to traditional methods. Then, we leverage an equivariant graph neural network, GATr, to perform track finding across all data from an event simultaneously. We validate the effectiveness of our pipeline on various detector concepts with different technologies for the FCC-ee at CERN, specifically the IDEA and CLD detectors. This work generalizes track finding across diverse types of input geometric data and tracking technologies, facilitating the development of innovative detector concepts, accelerating detector development cycles, and enabling comprehensive detector optimization.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 3055
Loading

OpenReview is a long-term project to advance science through improved peer review with legal nonprofit status. We gratefully acknowledge the support of the OpenReview Sponsors. © 2025 OpenReview