Neural network force fields for simple metals and semiconductors: construction and application to the calculation of phonons and melting temperatures

2019 ◽  
Vol 21 (12) ◽  
pp. 6506-6516 ◽  
Author(s):  
Mário R. G. Marques ◽  
Jakob Wolff ◽  
Conrad Steigemann ◽  
Miguel A. L. Marques
Keyword(s):  
Neural Network ◽  
Force Fields ◽  
Practical Procedure ◽  
Melting Temperatures

We present a practical procedure to obtain reliable and unbiased neural network based force fields for solids.

Friction Compensation in the Inverted Pendulum Controller by Means of a Neural Network

2020 ◽  
Vol 22 (4) ◽  
pp. 875-884
Author(s):  
Marek Balcerzak
Keyword(s):  
Neural Network ◽  
Inverted Pendulum ◽  
Friction Model ◽  
The Novel ◽  
Practical Procedure ◽  
The Neural Network ◽  
Wide Range ◽  
Novel Method ◽  
Data Acquisition And Processing ◽  
Friction Modelling

AbstractThis paper presents an experimental confirmation of the novel method of friction modelling and compensation. The method has been applied to an inverted pendulum control system. The practical procedure of data acquisition and processing has been described. Training of the neural network friction model has been covered. Application of the obtained model has been presented. The main asset of the presented model is its correctness in a wide range of relative velocities. Moreover, the model is relatively easy to build.

Profiling Molecular Simulations of SARS-CoV-2 Main Protease (Mpro) Binding to Repurposed Drugs Using Neural Network Force Fields

2020 ◽  
Author(s):  
Aayush Gupta
Keyword(s):  
Neural Network ◽  
Protein Interactions ◽  
Force Fields ◽  
Md Simulations ◽  
Interaction Sites ◽  
Main Protease ◽  
Rational Drug ◽  
Interaction Map ◽  
Novel Coronavirus ◽  
Approved Drugs

<div> <p> </p><div> <div> <div> <p> </p><div> <div> <div> <p> </p><div> <div> <div> <p>With the current pandemic situation caused by a novel coronavirus disease (COVID-19), there is an urgent call to develop a working therapeutic against it. Efficient computations aid to minimize the efforts by identifying a subset of drugs that can potentially bind to COVID-19 main protease or target protein (M<sup>PRO</sup>). The results of computations are always accompanied by their accuracy which depends on the details described by the model used. Machine learning models trained on millions of points and with unmatched accuracies are the best bet to employ in the process. In this work, I first identified and described the interaction sites of M<sup>PRO</sup> protein using a geometric deep learning model. Secondly, I conducted virtual screening (at one of the sites identified) on FDA approved drugs and picked 91 drugs having the highest binding affinity (below -8.0 kcal/mol). Then, I carried out 10 ns of molecular dynamics (MD) simulations using classical force fields and classified 37 drugs to be binding (includes drugs like Lopinavir, Saquinavir, Indinavir etc.) based on RMSD between MD-binding trajectories. To drastically improve the dynamics profile of selected 37 drugs, I brought in the highly accurate neural network force field (ANI) trained on coupled-cluster methods (CCSD(T)) data points and performed 1 ns of binding dynamics of each drug with protein. With the accurate approach, 19 drugs were qualified based on their RMSD cutoffs, and again with their free energy (ANI/MM/PBSA) computations another 7 drugs were rejected. The final selection of 12 drugs was validated based on MD trajectory clustering approach where 11 of 12 drugs (Targretin, Eltrombopag, Rifaximin, Deflazacort, Ergotamine, Doxazosin, Lastacaft, Rifampicin, Victrelis, Trajenta, Toposar, Indinavir) were confirmed to be binding. Further investigations were made to study their interactions with the protein and an accurate 2D- interaction map was generated. These findings and mapping of drug-protein interactions are highly accurate and could be potentially used to guide rational drug discovery against the COVID-19. </p> </div> </div> </div> </div> </div> </div> </div> </div> </div> </div>

Neural Network Force Fields for Metal Growth Based on Energy Decompositions

2020 ◽  
Vol 11 (4) ◽  
pp. 1364-1369 ◽  
Author(s):  
Qin Hu ◽  
Mouyi Weng ◽  
Xin Chen ◽  
Shucheng Li ◽  
Feng Pan ◽  
...  
Keyword(s):  
Neural Network ◽  
Force Fields

scholarly journals Neural network reactive force field for C, H, N, and O systems

2021 ◽  
Vol 7 (1) ◽  
Author(s):  
Pilsun Yoo ◽  
Michael Sakano ◽  
Saaketh Desai ◽  
Md Mahbubul Islam ◽  
Peilin Liao ◽  
...  
Keyword(s):  
Neural Network ◽  
Force Field ◽  
Density Functional ◽  
Force Fields ◽  
High Energy ◽  
Reactive Force ◽  
Dispersion Interactions ◽  
Reactive Force Field ◽  
Wide Range ◽  
Reactive Force Fields

AbstractReactive force fields have enabled an atomic level description of a wide range of phenomena, from chemistry at extreme conditions to the operation of electrochemical devices and catalysis. While significant insight and semi-quantitative understanding have been drawn from such work, the accuracy of reactive force fields limits quantitative predictions. We developed a neural network reactive force field (NNRF) for CHNO systems to describe the decomposition and reaction of the high-energy nitramine 1,3,5-trinitroperhydro-1,3,5-triazine (RDX). NNRF was trained using energies and forces of a total of 3100 molecules (11,941 geometries) and 15 condensed matter systems (32,973 geometries) obtained from density functional theory calculations with semi-empirical corrections to dispersion interactions. The training set is generated via a semi-automated iterative procedure that enables refinement of the NNRF until a desired accuracy is attained. The root mean square (RMS) error of NNRF on a testing set of configurations describing the reaction of RDX is one order of magnitude lower than current state of the art potentials.

Sign in / Sign up

Export Citation Format

Share Document