Super-resolution energy spectra from neutron direct-geometry spectrometers

Islam, Fahima; Lin, Jiao; Archibald, Rick; Abernathy, D. L.; Al-Qasir, I. I.; Campbell, Anne A.; Stone, M. B.; Granroth, G. E.

doi:10.1063/1.5116147

Cited by 10 publications

(4 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MCViNE has been improved in the last four years in software infrastructure and interfaces for users. Applications of the MCViNE package at SNS include assisting design of new instruments at the second target station, design of sample environments and collimators, as well as advanced data analysis [9,11,26,27]. MCViNE is an ongoing project.…”

Section: Discussionmentioning

confidence: 99%

Recent developments of MCViNE and its applications at SNS

et al. 2019

Self Cite

View full text Add to dashboard Cite

MCViNE is an open source, object-oriented Monte Carlo neutron ray-tracing simulation software package. Its design allows for flexible, hierarchical representations of sophisticated instrument components such as detector systems, and samples with a variety of shapes and scattering kernels. Recently this flexible design has enabled several applications of MCViNE simulations at the Spallation Neutron Source (SNS) at Oak Ridge National Lab, including assisting design of neutron instruments at the second target station and design of novel sample environments, as well as studying effects of instrument resolution and multiple scattering. Here we provide an overview of the recent developments and new features of MCViNE since its initial introduction (Jiao et al 2016 Nucl. Instrum. Methods Phys. Res., Sect. A 810, 86-99), and some example applications.

show abstract

Section: Discussionmentioning

confidence: 99%

Recent developments of MCViNE and its applications at SNS

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…Experiments at large-scale facilities such as neutron sources would seem to be a natural fit for the application of data-driven methods for analysis [7]. Indeed there have been a number of recent publications applying ML to many aspects of neutron science, from pulse discrimination in scintillator detectors [8], to enhancing resolution [9] and exploration of collected data [10]. ML techniques have been used for the analysis of diffuse neutron scattering in spin-ice systems [11], small angle neutron scattering [12,13] and for constructing inter-atomic potentials to compare to inelastic neutron scattering (INS) data [14].…”

Section: Introductionmentioning

confidence: 99%

Interpretable, calibrated neural networks for analysis and understanding of inelastic neutron scattering data

Butler

Thiyagalingam

et al. 2021

J. Phys.: Condens. Matter

View full text Add to dashboard Cite

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.

show abstract

“…Experiments at large-scale facilities such as neutron sources would seem to be a natural fit for the application of data-driven methods for analysis [7]. Indeed there have been a number of recent publications applying machine learning to many aspects of neutron science, from pulse discrimination in scintillator detectors [8], to enhancing resolution [9] and exploration of collected data [10]. ML techniques have been used for the analysis of diffuse neutron scattering in spin-ice systems [11], small angle neutron scattering [12,13] and for constructing inter-atomic potentials to compare to inelastic neutron scattering data [14].…”

Section: Introductionmentioning

confidence: 99%

Interpretable, calibrated neural networks for analysis and understanding of inelastic neutron scattering data

Butler,

Le,

Thiyagalingam

et al. 2020

Preprint

View full text Add to dashboard Cite

Deep neural networks 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 neural networks. In this work we examine approaches to all three issues. We use simulated data to train a deep neural network 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.

show abstract

Super-resolution energy spectra from neutron direct-geometry spectrometers

Cited by 10 publications

References 63 publications

Recent developments of MCViNE and its applications at SNS

Recent developments of MCViNE and its applications at SNS

Interpretable, calibrated neural networks for analysis and understanding of inelastic neutron scattering data

Interpretable, calibrated neural networks for analysis and understanding of inelastic neutron scattering data

Contact Info

Product

Resources

About