We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

High Energy Physics - Experiment

Title: Partition Pooling for Convolutional Graph Network Applications in Particle Physics

Authors: M. Bachlechner (1), T. Birkenfeld (1), P. Soldin (1), A. Stahl (1), C. Wiebusch (1) ((1) III Physics Institute B, RWTH Aachen University)
Abstract: Convolutional graph networks are used in particle physics for effective event reconstructions and classifications. However, their performances can be limited by the considerable amount of sensors used in modern particle detectors if applied to sensor-level data. We present a pooling scheme that uses partitioning to create pooling kernels on graphs, similar to pooling on images. Partition pooling can be used to adopt successful image recognition architectures for graph neural network applications in particle physics. The reduced computational resources allow for deeper networks and more extensive hyperparameter optimizations. To show its applicability, we construct a convolutional graph network with partition pooling that reconstructs simulated interaction vertices for an idealized neutrino detector. The pooling network yields improved performance and is less susceptible to overfitting than a similar network without pooling. The lower resource requirements allow the construction of a deeper network with further improved performance.
Subjects: High Energy Physics - Experiment (hep-ex); Machine Learning (cs.LG)
Journal reference: J. Instrum. 17, P10004 (2022)
DOI: 10.1088/1748-0221/17/10/P10004
Cite as: arXiv:2208.05952 [hep-ex]
  (or arXiv:2208.05952v2 [hep-ex] for this version)

Submission history

From: Thilo Birkenfeld [view email]
[v1] Thu, 11 Aug 2022 17:43:52 GMT (5311kb,D)
[v2] Mon, 5 Sep 2022 12:32:11 GMT (5289kb,D)

Link back to: arXiv, form interface, contact.