Integration of machine learning with neutron scattering for the Hamiltonian tuning of spin ice under pressure

Anjana Samarakoon; D. Alan Tennant; Feng Ye; Qiang Zhang; Santiago A. Grigera

doi:10.1038/s43246-022-00306-7

Integration of machine learning with neutron scattering for the Hamiltonian tuning of spin ice under pressure

Anjana Samarakoon, D. Alan Tennant, Feng Ye, Qiang Zhang, Santiago A. Grigera

Research output: Contribution to journal › Article › peer-review

9 Scopus citations

Abstract

Quantum materials research requires co-design of theory with experiments and involves demanding simulations and the analysis of vast quantities of data, usually including pattern recognition and clustering. Artificial intelligence is a natural route to optimise these processes and bring theory and experiments together. Here, we propose a scheme that integrates machine learning with high-performance simulations and scattering measurements, covering the pipeline of typical neutron experiments. Our approach uses nonlinear autoencoders trained on realistic simulations along with a fast surrogate for the calculation of scattering in the form of a generative model. We demonstrate this approach in a highly frustrated magnet, Dy₂Ti₂O₇, using machine learning predictions to guide the neutron scattering experiment under hydrostatic pressure, extract material parameters and construct a phase diagram. Our scheme provides a comprehensive set of capabilities that allows direct integration of theory along with automated data processing and provides on a rapid timescale direct insight into a challenging condensed matter system.

Original language	English
Article number	84
Journal	Communications Materials
Volume	3
Issue number	1
DOIs	https://doi.org/10.1038/s43246-022-00306-7
State	Published - Dec 2022

Funding

We acknowledge useful discussions on machine learning and workflows with Mingda Li, Tess Smidt, Scott Klasky, Juan Restrepo, Cristian Batista, and Guannan Zhang; and on glass formation in spin ice with Roderich Moessner and Claudio Castelnovo. A portion of this research used resources at the Spallation Neutron Source. The research by D.A.T. was sponsored by the Quantum Science Center. A.M.S. was supported by the U.S. Department of Energy, Office of Science, Materials Sciences and Engineering Division and Scientific User Facilities Division. S.A.G. acknowledges support from Agencia Nacional de Promoción Científica y Tecnológica through PICT 2017-2347. The computer modeling used resources of the Oak Ridge Leadership Computing Facility, which is supported by the Office of Science of the U.S. Department of Energy under contract no. DE-AC05-00OR22725.

Funders	Funder number
Materials Sciences and Engineering Division and Scientific User Facilities Division
Quantum Science Center
U.S. Department of Energy
Office of Science
Agencia Nacional de Promoción Científica y Tecnológica	DE-AC05-00OR22725, PICT 2017-2347

Access to Document

10.1038/s43246-022-00306-7

Cite this

@article{1decc42a993142d4ba366c43653ec277,

title = "Integration of machine learning with neutron scattering for the Hamiltonian tuning of spin ice under pressure",

abstract = "Quantum materials research requires co-design of theory with experiments and involves demanding simulations and the analysis of vast quantities of data, usually including pattern recognition and clustering. Artificial intelligence is a natural route to optimise these processes and bring theory and experiments together. Here, we propose a scheme that integrates machine learning with high-performance simulations and scattering measurements, covering the pipeline of typical neutron experiments. Our approach uses nonlinear autoencoders trained on realistic simulations along with a fast surrogate for the calculation of scattering in the form of a generative model. We demonstrate this approach in a highly frustrated magnet, Dy2Ti2O7, using machine learning predictions to guide the neutron scattering experiment under hydrostatic pressure, extract material parameters and construct a phase diagram. Our scheme provides a comprehensive set of capabilities that allows direct integration of theory along with automated data processing and provides on a rapid timescale direct insight into a challenging condensed matter system.",

author = "Anjana Samarakoon and Tennant, {D. Alan} and Feng Ye and Qiang Zhang and Grigera, {Santiago A.}",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s).",

year = "2022",

month = dec,

doi = "10.1038/s43246-022-00306-7",

language = "English",

volume = "3",

journal = "Communications Materials",

issn = "2662-4443",

publisher = "Nature Research",

number = "1",

}

TY - JOUR

T1 - Integration of machine learning with neutron scattering for the Hamiltonian tuning of spin ice under pressure

AU - Samarakoon, Anjana

AU - Tennant, D. Alan

AU - Ye, Feng

AU - Zhang, Qiang

AU - Grigera, Santiago A.

PY - 2022/12

Y1 - 2022/12

N2 - Quantum materials research requires co-design of theory with experiments and involves demanding simulations and the analysis of vast quantities of data, usually including pattern recognition and clustering. Artificial intelligence is a natural route to optimise these processes and bring theory and experiments together. Here, we propose a scheme that integrates machine learning with high-performance simulations and scattering measurements, covering the pipeline of typical neutron experiments. Our approach uses nonlinear autoencoders trained on realistic simulations along with a fast surrogate for the calculation of scattering in the form of a generative model. We demonstrate this approach in a highly frustrated magnet, Dy2Ti2O7, using machine learning predictions to guide the neutron scattering experiment under hydrostatic pressure, extract material parameters and construct a phase diagram. Our scheme provides a comprehensive set of capabilities that allows direct integration of theory along with automated data processing and provides on a rapid timescale direct insight into a challenging condensed matter system.

AB - Quantum materials research requires co-design of theory with experiments and involves demanding simulations and the analysis of vast quantities of data, usually including pattern recognition and clustering. Artificial intelligence is a natural route to optimise these processes and bring theory and experiments together. Here, we propose a scheme that integrates machine learning with high-performance simulations and scattering measurements, covering the pipeline of typical neutron experiments. Our approach uses nonlinear autoencoders trained on realistic simulations along with a fast surrogate for the calculation of scattering in the form of a generative model. We demonstrate this approach in a highly frustrated magnet, Dy2Ti2O7, using machine learning predictions to guide the neutron scattering experiment under hydrostatic pressure, extract material parameters and construct a phase diagram. Our scheme provides a comprehensive set of capabilities that allows direct integration of theory along with automated data processing and provides on a rapid timescale direct insight into a challenging condensed matter system.

UR - http://www.scopus.com/inward/record.url?scp=85141179354&partnerID=8YFLogxK

U2 - 10.1038/s43246-022-00306-7

DO - 10.1038/s43246-022-00306-7

M3 - Article

AN - SCOPUS:85141179354

SN - 2662-4443

VL - 3

JO - Communications Materials

JF - Communications Materials

IS - 1

M1 - 84

ER -

Integration of machine learning with neutron scattering for the Hamiltonian tuning of spin ice under pressure

Abstract

Funding

Access to Document

Other files and links

Fingerprint

Cite this