Ipp, A., Müller, D., Schuh, D., & Favoni, M. (2023, June 27). Visualizing the inner workings of L-CNNs [Conference Presentation]. Machine learning for lattice field theory and beyond 2023, Trento, Italy.
Machine learning for lattice field theory and beyond 2023
en
Event date:
26-Jun-2023 - 30-Jun-2023
-
Event place:
Trento, Italy
-
Keywords:
Convolutional Neural Networks; Lattice Gauge Theory
en
Abstract:
Lattice Gauge Equivariant Convolutional Neural Networks (L-CNNs) leverage convolutions with proper parallel transport and bilinear layers to combine basic plaquettes into arbitrarily shaped Wilson loops of growing length and area [1]. These networks provide a powerful framework for addressing challenging problems in lattice field theory.
In this talk, we explore the inner workings of L-CNNs, aiming to gain insight into the contributions of the different layers. Through visualization techniques, we analyze the patterns and structures of the Wilson loops that emerge, studying to what degree L-CNN architectures exhibit redundancy in the parameters. With our findings we aim to provide a deeper understanding of L-CNN behavior and improve its interpretability.
en
Project title:
Glasma-Simulationen mit maschinellem Lernen hochskalieren: P 32446-N27 (FWF - Österr. Wissenschaftsfonds) Simulation der frühesten Stadien von Schwerionenkollisionen: P 34764-N (FWF - Österr. Wissenschaftsfonds) Nichtperturbative Eigenschaften evolvierenden gluonischer Plasmen: P 34455-N (FWF - Österr. Wissenschaftsfonds)
