Machine-learned approximations to Density Functional Theory Hamiltonians

Large scale Density Functional Theory (DFT) based electronic structure calculations are highly time consuming and scale poorly with system size. While semi-empirical approximations to DFT result in a reduction in computational time versus ab initio DFT, creating such approximations involves significant manual intervention and is highly inefficient for high-throughput electronic structure screening calculations. In this letter, we propose the use of machine-learning for prediction of DFT Hamiltonians. Using suitable representations of atomic neighborhoods and Kernel Ridge Regression, we show that an accurate and transferable prediction of DFT Hamiltonians for a variety of material environments can be achieved. Electronic structure properties such as ballistic transmission and band structure computed using predicted Hamiltonians compare accurately with their DFT counterparts. The method is independent of the specifics of the DFT basis or material system used and can easily be automated and scaled for predicting Hamiltonians of any material system of interest.

[1]  D. Pettifor,et al.  New many-body potential for the bond order. , 1989, Physical review letters.

[2]  C. Wang,et al.  Environment-Dependent Tight-Binding Potential Models , 1997 .

[3]  David R. Bowler,et al.  Tight-binding modelling of materials , 1997 .

[4]  Sándor Suhai,et al.  Self-consistent-charge density-functional tight-binding method for simulations of complex materials properties , 1998 .

[5]  HOMAS,et al.  Si tight-binding parameters from genetic algorithm fitting , 2000 .

[6]  Trevor Hastie,et al.  The Elements of Statistical Learning , 2001 .

[7]  Bernhard Schölkopf,et al.  A tutorial on support vector regression , 2004, Stat. Comput..

[8]  P. Vogl,et al.  Compact expression for the angular dependence of tight-binding Hamiltonian matrix elements , 2004 .

[9]  Michele Parrinello,et al.  Generalized neural-network representation of high-dimensional potential-energy surfaces. , 2007, Physical review letters.

[10]  Tight-binding Hamiltonian from first-principles calculations , 2008 .

[11]  Wen-Cai Lu,et al.  Tight-binding Hamiltonian from first-principles calculations , 2008 .

[12]  Sidney Yip,et al.  Quasiatomic orbitals for ab initio tight-binding analysis , 2008 .

[13]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[14]  R. Kondor,et al.  Gaussian approximation potentials: the accuracy of quantum mechanics, without the electrons. , 2009, Physical review letters.

[15]  B. Meyer,et al.  Parameterization of tight-binding models from density functional theory calculations , 2011 .

[16]  Hsuan-Tien Lin,et al.  Learning From Data , 2012 .

[17]  Andreas Ziehe,et al.  Learning Invariant Representations of Molecules for Atomization Energy Prediction , 2012, NIPS.

[18]  Klaus-Robert Müller,et al.  Finding Density Functionals with Machine Learning , 2011, Physical review letters.

[19]  D. Bowler,et al.  O(N) methods in electronic structure calculations. , 2011, Reports on progress in physics. Physical Society.

[20]  K. Müller,et al.  Fast and accurate modeling of molecular atomization energies with machine learning. , 2011, Physical review letters.

[21]  M. Rupp,et al.  Fourier series of atomic radial distribution functions: A molecular fingerprint for machine learning models of quantum chemical properties , 2013, 1307.2918.

[22]  R. Kondor,et al.  On representing chemical environments , 2012, 1209.3140.

[23]  Risi Kondor,et al.  Publisher's Note: On representing chemical environments , 2013 .

[24]  T. Boykin,et al.  An environment-dependent semi-empirical tight binding model suitable for electron transport in bulk metals, metal alloys, metallic interfaces, and metallic nanostructures. I. Model and validation , 2013, 1311.6082.

[25]  Kristof T. Schütt,et al.  How to represent crystal structures for machine learning: Towards fast prediction of electronic properties , 2013, 1307.1266.

[26]  O. Anatole von Lilienfeld,et al.  Modeling electronic quantum transport with machine learning , 2014, 1401.8277.

[27]  K. Müller,et al.  Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space , 2015, The journal of physical chemistry letters.

[28]  Gábor Csányi,et al.  Gaussian approximation potentials: A brief tutorial introduction , 2015, 1502.01366.

[29]  Matthias Rupp,et al.  Machine learning for quantum mechanics in a nutshell , 2015 .

[30]  Christian Trott,et al.  Spectral neighbor analysis method for automated generation of quantum-accurate interatomic potentials , 2014, J. Comput. Phys..

[31]  Gerhard Klimeck,et al.  Tight-binding analysis of Si and GaAs ultrathin bodies with subatomic wave-function resolution , 2015, Physical Review B.

[32]  On the feasibility of ab initio electronic structure calculations for Cu using a single s orbital basis , 2015, 1509.01204.

[33]  Transferable tight-binding model for strained group IV and III-V materials and heterostructures , 2016, 1603.05266.

[34]  M. Rodder,et al.  Is electron transport in nanocrystalline Cu interconnects surface dominated or grain boundary dominated? , 2016, 2016 IEEE International Interconnect Technology Conference / Advanced Metallization Conference (IITC/AMC).

[35]  M. Rodder,et al.  Lower limits of line resistance in nanocrystalline back end of line Cu interconnects , 2016, 1601.06675.