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

Butler, Keith T.; Le, Manh Duc; Thiyagalingam, Jeyarajan; Perring, T. G.

doi:10.1088/1361-648x/abea1c

Cited by 19 publications

(42 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The regions identified by the CNN/CAM match with the regions that a trained physicist identifies, but in a fraction of the time. Spectra and maps reproduced under the CC-BY 4.0 License from ref .…”

Section: Model Explanation Methodsmentioning

confidence: 99%

Interpretable and Explainable Machine Learning for Materials Science and Chemistry

et al. 2022

Self Cite

View full text Add to dashboard Cite

Conspectus Machine learning has become a common and powerful tool in materials research. As more data become available, with the use of high-performance computing and high-throughput experimentation, machine learning has proven potential to accelerate scientific research and technology development. Though the uptake of data-driven approaches for materials science is at an exciting, early stage, to realize the true potential of machine learning models for successful scientific discovery, they must have qualities beyond purely predictive power. The predictions and inner workings of models should provide a certain degree of explainability by human experts, permitting the identification of potential model issues or limitations, building trust in model predictions, and unveiling unexpected correlations that may lead to scientific insights. In this work, we summarize applications of interpretability and explainability techniques for materials science and chemistry and discuss how these techniques can improve the outcome of scientific studies. We start by defining the fundamental concepts of interpretability and explainability in machine learning and making them less abstract by providing examples in the field. We show how interpretability in scientific machine learning has additional constraints compared to general applications. Building upon formal definitions in machine learning, we formulate the basic trade-offs among the explainability, completeness, and scientific validity of model explanations in scientific problems. In the context of these trade-offs, we discuss how interpretable models can be constructed, what insights they provide, and what drawbacks they have. We present numerous examples of the application of interpretable machine learning in a variety of experimental and simulation studies, encompassing first-principles calculations, physicochemical characterization, materials development, and integration into complex systems. We discuss the varied impacts and uses of interpretabiltiy in these cases according to the nature and constraints of the scientific study of interest. We discuss various challenges for interpretable machine learning in materials science and, more broadly, in scientific settings. In particular, we emphasize the risks of inferring causation or reaching generalization by purely interpreting machine learning models and the need for uncertainty estimates for model explanations. Finally, we showcase a number of exciting developments in other fields that could benefit interpretability in material science problems. Adding interpretability to a machine learning model often requires no more technical know-how than building the model itself. By providing concrete examples of studies (many with associated open source code and data), we hope that this Account will encourage all practitioners of machine learning in materials science to look deeper into their models.

show abstract

Section: Model Explanation Methodsmentioning

confidence: 99%

Interpretable and Explainable Machine Learning for Materials Science and Chemistry

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Spinwave theory provides fast computation for simple magnon dynamics and codes are available 29 . More interestingly, machine learning holds the promise of being able to interface sophisticated simulations 30 that include quantum effects including correlations which are important to quantum materials. Examples are density matrix renormalization group 31 , quantum Monte Carlo 32 , dynamical mean field theory 33 and dynamic cluster approximation approaches to name but a few.…”

Section: Machine Learningmentioning

confidence: 99%

Integration of Machine Learning with Neutron Scattering: Hamiltonian Tuning in Spin Ice with Pressure

Samarakoon¹,

Tennant²,

Ye³

et al. 2021

Preprint

View full text Add to dashboard Cite

FIG. 1. Structure of the magnetic system, phase map and generative model. (a) The magnetic moments located on Dy ions are constrained by crystal field interactions to point in or out of the tetrahedra. They form a corner sharing pyrochlore lattice. Nearest neighbors (1), next-nearest-neighbors (2) and two inequivalent next-next-nearest neighbors (3 and 3') interactions are shown as thick colored lines. (b) A predicted map of magnetic orderings for varying J3 and J 3 with the remaining Hamiltonian parameters J1, J2, and D fixed to 3.41 K. 0 K and 1.3224 K respectively. The latent space coordinates, S(L) predicted by the Generator network (GN) have been converted to an RGB color code. A region with uniform color is expected to be structurally the same, and continuous color changes correspond to either crossovers or continuous transitions. (c) Comparison between the high-symmetry-plane slices of simulated and surrogate-predicted S(Q) data at multiple places in parameter space as indexed on panel (b). (d) Schematic diagram of the experimental parameters that can be used to tune the properties of frustrated materials and their associated effects. (e) Schematic design of the surrogate model used to predict S(Q) for a given set of model parameters. The surrogate comprises a radial basis network, the Generator, mapping parameter space to latent space and a Decoder to reconstruct S(Q) from latent space representations.

show abstract

“…Machine-learning-based methods have recently been applied to experimental data in large experimental facilities (Hey et al, 2020), including neutron scattering data (Butler et al, 2021;Archibald et al, 2020;Demerdash et al, 2019;Samarakoon et al, 2020). The optimization of the histogram representation will provide a basic tool to proceed with machine-learning studies because the optimized histogram shows the right amount of information in the data in the context that it represents the underlying probability density most accurately among histograms with equal bin widths.…”

Section: Introductionmentioning

confidence: 99%

Optimization and inference of bin widths for histogramming inelastic neutron scattering spectra

Tatsumi¹,

Inamura²,

Kofu³

et al. 2022

J Appl Cryst

View full text Add to dashboard Cite

A data-driven bin-width optimization for the histograms of measured data sets based on inhomogeneous Poisson processes was developed in a neurophysiology study [Shimazaki & Shinomoto (2007). Neural Comput. 19, 1503–1527], and a subsequent study [Muto, Sakamoto, Matsuura, Arima & Okada (2019). J. Phys. Soc. Jpn, 88, 044002] proposed its application to inelastic neutron scattering (INS) data. In the present study, the results of the method on experimental INS time-of-flight data collected under different measurement conditions from a copper single crystal are validated. The extrapolation of the statistics on a given data set to other data sets with different total counts precisely infers the optimal bin widths on the latter. The histograms with the optimized bin widths statistically verify two fine-spectral-feature examples in the energy and momentum transfer cross sections: (i) the existence of phonon band gaps; and (ii) the number of plural phonon branches located close to each other. This indicates that the applied method helps in the efficient and rigorous observation of spectral structures important in physics and materials science like novel forms of magnetic excitation and phonon states correlated to thermal conductivities.

show abstract

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

Cited by 19 publications

References 43 publications

Interpretable and Explainable Machine Learning for Materials Science and Chemistry

Interpretable and Explainable Machine Learning for Materials Science and Chemistry

Integration of Machine Learning with Neutron Scattering: Hamiltonian Tuning in Spin Ice with Pressure

Optimization and inference of bin widths for histogramming inelastic neutron scattering spectra

Contact Info

Product

Resources

About