Dataset Information

Machine-learning-assisted insight into spin ice Dy₂Ti₂O₇.

ABSTRACT: Complex behavior poses challenges in extracting models from experiment. An example is spin liquid formation in frustrated magnets like Dy₂Ti₂O₇. Understanding has been hindered by issues including disorder, glass formation, and interpretation of scattering data. Here, we use an automated capability to extract model Hamiltonians from data, and to identify different magnetic regimes. This involves training an autoencoder to learn a compressed representation of three-dimensional diffuse scattering, over a wide range of spin Hamiltonians. The autoencoder finds optimal matches according to scattering and heat capacity data and provides confidence intervals. Validation tests indicate that our optimal Hamiltonian accurately predicts temperature and field dependence of both magnetic structure and magnetization, as well as glass formation and irreversibility in Dy₂Ti₂O₇. The autoencoder can also categorize different magnetic behaviors and eliminate background noise and artifacts in raw data. Our methodology is readily applicable to other materials and types of scattering problems.

SUBMITTER: Samarakoon AM

PROVIDER: S-EPMC7021707 | biostudies-literature | 2020 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine-learning-assisted insight into spin ice Dy2Ti2O7.

Samarakoon Anjana M AM Barros Kipton K Li Ying Wai YW Eisenbach Markus M Zhang Qiang Q Ye Feng F Sharma V V Dun Z L ZL Zhou Haidong H Grigera Santiago A SA Batista Cristian D CD Tennant D Alan DA

Nature communications 20200214 1

Complex behavior poses challenges in extracting models from experiment. An example is spin liquid formation in frustrated magnets like Dy2Ti2O7. Understanding has been hindered by issues including disorder, glass formation, and interpretation of scattering data. Here, we use an automated capability to extract model Hamiltonians from data, and to identify different magnetic regimes. This involves training an autoencoder to learn a compressed representation of thre ...[more]

PMID: 32060263

Similar Datasets

Project description:Molybdenum disulfide (MoS2) is a layered transition metal-sulfur compound semiconductor that shows promising prospects for applications in optoelectronics and integrated circuits because of its low preparation cost, good stability and excellent physicochemical, biological and mechanical properties. MoS2 with high quality, large size and outstanding performance can be prepared via chemical vapor deposition (CVD). However, its preparation process is complex, and the area of MoS2 obtained is difficult to control. Machine learning (ML), as a powerful tool, has been widely applied in materials science. Based on this, in this paper, a ML Gaussian regression model was constructed to explore the growth mechanism of MoS2 material prepared with the CVD method. The parameters of the regression model were evaluated by combining the four indicators of goodness of fit (r2), mean squared error (MSE), Pearson correlation coefficient (p) and p-value (p_val) of Pearson's correlation coefficient. After comprehensive comparison, it was found that the performance of the model was optimal when the number of iterations was 15. Additionally, feature importance analysis was conducted on the growth parameters using the established model. The results showed that the carrier gas flow rate (Fr), molybdenum sulfur ratio (R) and reaction temperature (T) had a crucial impact on the CVD growth of MoS2 materials. The optimal model was used to predict the size of molybdenum disulfide synthesis under 185,900 experimental conditions in the simulation dataset so as to select the optimal range for the synthesis of large-size molybdenum disulfide. Furthermore, the model prediction results were verified through literature and experimental results. It was found that the relative error between the prediction results and the literature and experimental results was small. These findings provide an effective solution to the preparation of MoS2 materials with a reduction in the time and cost of trial and error.

Project description:The ecological and environmental science communities have embraced machine learning (ML) for empirical modelling and prediction. However, going beyond prediction to draw insights into underlying functional relationships between response variables and environmental 'drivers' is less straightforward. Deriving ecological insights from fitted ML models requires techniques to extract the 'learning' hidden in the ML models.We revisit the theoretical background and effectiveness of four approaches for deriving insights from ML: ranking independent variable importance (Gini importance, GI; permutation importance, PI; split importance, SI; and conditional permutation importance, CPI), and two approaches for inference of bivariate functional relationships (partial dependence plots, PDP; and accumulated local effect plots, ALE). We also explore the use of a surrogate model for visualization and interpretation of complex multi-variate relationships between response variables and environmental drivers. We examine the challenges and opportunities for extracting ecological insights with these interpretation approaches. Specifically, we aim to improve interpretation of ML models by investigating how effectiveness relates to (a) interpretation algorithm, (b) sample size and (c) the presence of spurious explanatory variables.We base the analysis on simulations with known underlying functional relationships between response and predictor variables, with added white noise and the presence of correlated but non-influential variables. The results indicate that deriving ecological insight is strongly affected by interpretation algorithm and spurious variables, and moderately impacted by sample size. Removing spurious variables improves interpretation of ML models. Meanwhile, increasing sample size has limited value in the presence of spurious variables, but increasing sample size does improves performance once spurious variables are omitted. Among the four ranking methods, SI is slightly more effective than the other methods in the presence of spurious variables, while GI and SI yield higher accuracy when spurious variables are removed. PDP is more effective in retrieving underlying functional relationships than ALE, but its reliability declines sharply in the presence of spurious variables. Visualization and interpretation of the interactive effects of predictors and the response variable can be enhanced using surrogate models, including three-dimensional visualizations and use of loess planes to represent independent variable effects and interactions.Machine learning analysts should be aware that including correlated independent variables in ML models with no clear causal relationship to response variables can interfere with ecological inference. When ecological inference is important, ML models should be constructed with independent variables that have clear causal effects on response variables. While interpreting ML models for ecological inference remains challenging, we show that careful choice of interpretation methods, exclusion of spurious variables and adequate sample size can provide more and better opportunities to 'learn from machine learning'.

Dataset Information

Machine-learning-assisted insight into spin ice Dy₂Ti₂O₇.

Publications

Machine-learning-assisted insight into spin ice Dy<sub>2</sub>Ti<sub>2</sub>O<sub>7</sub>.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets