. (2015) A framework for machine-learning-augmented multiscale atomistic simulations on parallel supercomputers. International Journal of Quantum Chemistry. Permanent WRAP url: http://wrap.warwick.ac.uk/68012 Copyright and reuse:The Warwick Research Archive Portal (WRAP) makes this work of researchers of the University of Warwick available open access under the following conditions. Copyright © and all moral rights to the version of the paper presented here belong to the individual author(s) and/or other copyright owners. To the extent reasonable and practicable the material made available in WRAP has been checked for eligibility before being made available.Copies of full items can be used for personal research or study, educational, or not-forprofit purposes without prior permission or charge. Provided that the authors, title and full bibliographic details are credited, a hyperlink and/or URL is given for the original metadata page and the content is not changed in any way. Publisher statement:"This is the peer reviewed version of the following article: Caccin, Marco, Li, Zhenwei, Kermode, James R and De Vita, Alessandro. (2015) A framework for machine-learningaugmented multiscale atomistic simulations on parallel supercomputers. International Journal of Quantum Chemistry, which has been published in final form at http://dx.doi.org/10.1002/qua.24952 . This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving." A note on versions:The version presented here may differ from the published version or, version of record, if you wish to cite this item you are advised to consult the publisher's version. Please see the 'permanent WRAP url' above for details on accessing the published version and note that access may require a subscription. AbstractRecent advances in quantum mechanical(QM)-based molecular dynamics simulations have used machine-learning (ML) to predict, rather than re-calculate, QMaccurate forces in atomic configurations sufficiently similar to previously encountered ones. Here, we discuss how ML approaches can be deployed within large-scale QM/MM materials simulations on massively parallel supercomputers, making QM zones of 1000 atoms routinely attainable. We argue that the ML approach allows computational effort to be concentrated on the most chemically active subregions of the QM zone, significantly improving the overall efficiency of the simulation. We thus propose a novel method to partition large QM regions into multiple subregions which can be computed in parallel to achieve optimal scaling. Then we review a recently proposed QM/ML MD scheme [Z. Li et al., Phys. Rev. Lett. 114(9), 096405 (2015)], discussing how this could be efficiently combined with QM-zone partitioning.

No abstract

The Tan‐Lu Fault Zone is a major fault system in eastern China, and its timing of activity has been the focus of much research. We studied a ductile segment of the Tan‐Lu Fault Zone, the Malongshan Shear Zone, which is located in the Feidong Complex (the southern part of the Zhangbaling Uplift). A dark biotite‐adamellite mylonite has been recognized in the Malongshan Shear Zone, and the mylonitic foliation strikes NNE, consistent with the regional orientation of the Tan‐Lu Fault Zone. The occurrence of abundant syn‐tectonic leucocratic veins within the biotite‐adamellite mylonite indicates that ductile shearing was accompanied by hydrothermal activity. To constrain the timing of deformation along the Malongshan Shear Zone, we dated the biotite‐adamellite mylonite and the leucocratic veins using zircon U–Pb geochronology. The weighted mean ages of zircon grains from the biotite‐adamellite mylonite (159.0 ± 5.5 Ma) and the captured Group 2 (156.3 ± 3.7 Ma and 173.2 ± 4.2 Ma) of leucocratic veins are interpreted as the ages of protoliths of the biotite‐adamellite mylonite during the Jurassic. Additional weighted mean age of the hydrothermal zircon grains (130.9 ± 3.6 Ma) represents the timing of crystallization of the veins. Therefore, the results indicate that the Malongshan Shear Zone (and hence the southern segment of the Tan‐Lu Fault Zone) was active during the Early Cretaceous (~131 Ma). Based on the present results and existing data, we consider that sinistral shearing of the Tan‐Lu Fault Zone during the Early Cretaceous was controlled by oblique subduction of the Pacific Plate at a high angle.

Through empirical research, it is found that the traditional autoregressive integrated moving average (ARIMA) model has a large deviation for the forecasting of high-frequency financial time series. With the improvement in storage capacity and computing power of high-frequency financial time series, this paper combines the traditional ARIMA model with the deep learning model to forecast high-frequency financial time series. It not only preserves the theoretical basis of the traditional model and characterizes the linear relationship, but also can characterize the nonlinear relationship of the error term according to the deep learning model. The empirical study of Monte Carlo numerical simulation and CSI 300 index in China show that, compared with ARIMA, support vector machine (SVM), long short-term memory (LSTM) and ARIMA-SVM models, the improved ARIMA model based on LSTM not only improves the forecasting accuracy of the single ARIMA model in both fitting and forecasting, but also reduces the computational complexity of only a single deep learning model. The improved ARIMA model based on deep learning not only enriches the models for the forecasting of time series, but also provides effective tools for high-frequency strategy design to reduce the investment risks of stock index.