Regensburg 2022 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
DY: Fachverband Dynamik und Statistische Physik
DY 45: Poster Session: Nonlinear Dynamics, Pattern Formation, Data Analytics and Machine Learning
DY 45.8: Poster
Donnerstag, 8. September 2022, 15:00–18:00, P2
Neural Network-Based Approaches for Multiscale Modelling of Topological Defects — •Kyra Klos1, Karin Everschor-Sitte2, and Friederike Schmid1 — 1Johannes Gutenberg University, Mainz, Germany — 2University of Duisburg-Essen, Duisburg, Germany
Topological defects and their dynamics are a heavily researched topic in a wide range of physics fields [1].
Due to the multiscale character of those defect structures, numerically simulating a large number of them in full microscopic detail gets highly computationally expensive, as the large size of associated deformation fields around each core leads to a complex interaction pattern.
To give a possible insight into the connection between the macroscopic (particle) description of a model with topological defects and the underlying microscopic structure, we propose the use of neural networks. With a spin-dynamic simulated microscopic model as training data, we use a conditional generative adversarial network system [2] with Wasserstein-loss [3] to generate reasonable spin-configurations from given defect configuration inputs. To guarantee the generation of realistic spin configuration, we also include additional physical constraints into our generator.
[1] Mermin N. D., Rev. Mod. Phys. 51, 591, (1979)
[2] Mirza M. ; Osindero S., arXiv:1411.1784v1, (2014)
[3] Arjovsky M. et al., ICML, PMLR 70, 214, (2017)
[4] Goodfellow I. et al., NeurIPS, (2014)