The sheer number of parameters in deep learning makes the physical interpretation of failure predictions in glasses challenging. Here the authors use Grad-CAM to reveal the role of topological defects and local potential energies in failure predictions.Being able to predict the failure of materials based on structural information is a fundamental issue with enormous practical and industrial relevance for the monitoring of devices and components. Thanks to recent advances in deep learning, accurate failure predictions are becoming possible even for strongly disordered solids, but the sheer number of parameters used in the process renders a physical interpretation of the results impossible. Here we address this issue and use machine learning methods to predict the failure of simulated two dimensional silica glasses from their initial undeformed structure. We then exploit Gradient-weighted Class Activation Mapping (Grad-CAM) to build attention maps associated with the predictions, and we demonstrate that these maps are amenable to physical interpretation in terms of topological defects and local potential energies. We show that our predictions can be transferred to samples with different shape or size than those used in training, as well as to experimental images. Our strategy illustrates how artificial neural networks trained with numerical simulation results can provide interpretable predictions of the behavior of experimentally measured structures.

Predicting the failure of two-dimensional silica glasses / F. FONT CLOS, M. Zanchi, S. Hiemer, S. Bonfanti, R. Guerra, M. Zaiser, S. Zapperi. - In: NATURE COMMUNICATIONS. - ISSN 2041-1723. - 13:1(2022 May 20), pp. 2820.1-2820.11. [10.1038/s41467-022-30530-1]

Predicting the failure of two-dimensional silica glasses

F. FONT CLOS
Primo
;
M. Zanchi
Secondo
;
S. Bonfanti;R. Guerra;S. Zapperi
Ultimo
2022

Abstract

The sheer number of parameters in deep learning makes the physical interpretation of failure predictions in glasses challenging. Here the authors use Grad-CAM to reveal the role of topological defects and local potential energies in failure predictions.Being able to predict the failure of materials based on structural information is a fundamental issue with enormous practical and industrial relevance for the monitoring of devices and components. Thanks to recent advances in deep learning, accurate failure predictions are becoming possible even for strongly disordered solids, but the sheer number of parameters used in the process renders a physical interpretation of the results impossible. Here we address this issue and use machine learning methods to predict the failure of simulated two dimensional silica glasses from their initial undeformed structure. We then exploit Gradient-weighted Class Activation Mapping (Grad-CAM) to build attention maps associated with the predictions, and we demonstrate that these maps are amenable to physical interpretation in terms of topological defects and local potential energies. We show that our predictions can be transferred to samples with different shape or size than those used in training, as well as to experimental images. Our strategy illustrates how artificial neural networks trained with numerical simulation results can provide interpretable predictions of the behavior of experimentally measured structures.
Settore FIS/03 - Fisica della Materia
20-mag-2022
Article (author)
File in questo prodotto:
File Dimensione Formato  
Font-Clos-Ncomms2022.pdf

accesso aperto

Tipologia: Publisher's version/PDF
Dimensione 7.96 MB
Formato Adobe PDF
7.96 MB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2434/941009
Citazioni
  • ???jsp.display-item.citation.pmc??? 3
  • Scopus 11
  • ???jsp.display-item.citation.isi??? 8
social impact