Abstract
Deep neural networks (NNs) provide flexible frameworks for learning data representations and functions relating data to other properties and are often claimed to achieve ‘super-human’ performance in inferring relationships between input data and desired property. In the context of inelastic neutron scattering experiments, however, as in many other scientific scenarios, a number of issues arise: (i) scarcity of labelled experimental data, (ii) lack of uncertainty quantification on results, and (iii) lack of interpretability of the deep NNs. In this work we examine approaches to all three issues. We use simulated data to train a deep NN to distinguish between two possible magnetic exchange models of a half-doped manganite. We apply the recently developed deterministic uncertainty quantification method to provide error estimates for the classification, demonstrating in the process how important realistic representations of instrument resolution in the training data are for reliable estimates on experimental data. Finally we use class activation maps to determine which regions of the spectra are most important for the final classification result reached by the network.
Highlights
Neural networks have witnessed a renaissance in the last decade, dramatically improving on existing state of the art performance in fields from image processing to automatic language translation
It has been proposed that we are on the cusp of a “fourth paradigm” of data-driven discovery [1, 2], an idea that is supported by the recent explosion in research using machine learning (ML), not least in materials science and condensed matter physics [3, 4, 5, 6]
While these shifts do not pose a problem for the training of the network which is done on generated data, it may become an issue when confronted by experimental data, because slight miscalibrations of the spectrometer may produce data which is offset compared to the ideal simulations
Summary
Neural networks have witnessed a renaissance in the last decade, dramatically improving on existing state of the art performance in fields from image processing to automatic language translation. In the case of quasi-2D magnetic materials, where the magnetic coupling is significant only within planes of atoms but negligible between planes, these three coordinates can be transformed into those of wave vector within the plane and energy transfer, thereby in principle enabling the entire excitation spectrum to be measured in parallel An extension of this approach is to combine ∼ 100 − 500 measurements between which the crystal is rotated by a few tenths of a degree. The motivation for this work has been to investigate to what extent ML methods can reduce the labour of inelastic neutron data analysis while still returning reliable and interpretable information In this contribution we re-examine the data from a well-understood INS measurement of spin waves in a single crystal of Pr(Ca0.9Sr0.1)2Mn2O7 [26], a moderately complex magnetic material, using recent advances in deep learning.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
