Parameterization of empirical forcefields for glassy silica using machine learning
- PDF / 1,191,912 Bytes
- 7 Pages / 612 x 792 pts (letter) Page_size
- 109 Downloads / 143 Views
rtificial Intelligence Research Letter
Parameterization of empirical forcefields for glassy silica using machine learning Han Liu, Physics of AmoRphous and Inorganic Solids Laboratory (PARISlab), Department of Civil and Environmental Engineering, University of California, Los Angeles, CA 90095, USA Zipeng Fu, Physics of AmoRphous and Inorganic Solids Laboratory (PARISlab), Department of Civil and Environmental Engineering, University of California, Los Angeles, CA 90095, USA; Department of Computer Science, University of California, Los Angeles, CA 90095, USA Yipeng Li, Nazreen Farina Ahmad Sabri, and Mathieu Bauchy, Physics of AmoRphous and Inorganic Solids Laboratory (PARISlab), Department of Civil and Environmental Engineering, University of California, Los Angeles, CA 90095, USA Address all correspondence to Mathieu Bauchy at [email protected] (Received 14 January 2019; accepted 29 March 2019)
Abstract The development of reliable, yet computationally efficient interatomic forcefields is key to facilitate the modeling of glasses. However, the parameterization of novel forcefields is challenging as the high number of parameters renders traditional optimization methods inefficient or subject to bias. Here, we present a new parameterization method based on machine learning, which combines ab initio molecular dynamics simulations and Bayesian optimization. By taking the example of glassy silica, we show that our method yields a new interatomic forcefield that offers an unprecedented agreement with ab initio simulations. This method offers a new route to efficiently parameterize new interatomic forcefields for disordered solids in a non-biased fashion.
Introduction Classical molecular dynamics (MD) simulation is an effective tool to access the atomic structure of glass, which usually remains invisible from traditional experimental techniques.[1–3] In turn, better understanding the atomic structure of glasses is key to decipher their genome, that is, to understand how their composition and structure control their engineering properties.[4] However, the accuracy of glass modeling based on MD simulations largely depends on the reliability of the underlying empirical forcefield, i.e., two-body (and sometimes threebody or more) interatomic potential.[3,5] Although ab initio molecular dynamics (AIMD) can, in theory, overcome these limitations, the high-computational cost of this technique renders challenging glass simulations—which typically require large systems for statistical averaging and long timescales to slowly quench a melt down to the glassy state.[3,6,7] The development of new, improved empirical forcefields presently represents a bottleneck in glass modeling.[8–10] Empirical forcefields are typically based on functionals that depend on several parameters (e.g., partial atomic charges, etc.), which need to be properly optimized in order to minimize a given cost function.[11,12] One option is to define the cost function in terms of the difference between the structure or properties of the simulated system and available experime
Data Loading...
