-
Grad DFT: a software library for machine learning enhanced density functional theory
Authors:
Pablo A. M. Casares,
Jack S. Baker,
Matija Medvidovic,
Roberto dos Reis,
Juan Miguel Arrazola
Abstract:
Density functional theory (DFT) stands as a cornerstone method in computational quantum chemistry and materials science due to its remarkable versatility and scalability. Yet, it suffers from limitations in accuracy, particularly when dealing with strongly correlated systems. To address these shortcomings, recent work has begun to explore how machine learning can expand the capabilities of DFT; an…
▽ More
Density functional theory (DFT) stands as a cornerstone method in computational quantum chemistry and materials science due to its remarkable versatility and scalability. Yet, it suffers from limitations in accuracy, particularly when dealing with strongly correlated systems. To address these shortcomings, recent work has begun to explore how machine learning can expand the capabilities of DFT; an endeavor with many open questions and technical challenges. In this work, we present Grad DFT: a fully differentiable JAX-based DFT library, enabling quick prototyping and experimentation with machine learning-enhanced exchange-correlation energy functionals. Grad DFT employs a pioneering parametrization of exchange-correlation functionals constructed using a weighted sum of energy densities, where the weights are determined using neural networks. Moreover, Grad DFT encompasses a comprehensive suite of auxiliary functions, notably featuring a just-in-time compilable and fully differentiable self-consistent iterative procedure. To support training and benchmarking efforts, we additionally compile a curated dataset of experimental dissociation energies of dimers, half of which contain transition metal atoms characterized by strong electronic correlations. The software library is tested against experimental results to study the generalization capabilities of a neural functional across potential energy surfaces and atomic species, as well as the effect of training data noise on the resulting model accuracy.
△ Less
Submitted 11 December, 2023; v1 submitted 22 September, 2023;
originally announced September 2023.
-
Parallel hybrid quantum-classical machine learning for kernelized time-series classification
Authors:
Jack S. Baker,
Gilchan Park,
Kwangmin Yu,
Ara Ghukasyan,
Oktay Goktas,
Santosh Kumar Radha
Abstract:
Supervised time-series classification garners widespread interest because of its applicability throughout a broad application domain including finance, astronomy, biosensors, and many others. In this work, we tackle this problem with hybrid quantum-classical machine learning, deducing pairwise temporal relationships between time-series instances using a time-series Hamiltonian kernel (TSHK). A TSH…
▽ More
Supervised time-series classification garners widespread interest because of its applicability throughout a broad application domain including finance, astronomy, biosensors, and many others. In this work, we tackle this problem with hybrid quantum-classical machine learning, deducing pairwise temporal relationships between time-series instances using a time-series Hamiltonian kernel (TSHK). A TSHK is constructed with a sum of inner products generated by quantum states evolved using a parameterized time evolution operator. This sum is then optimally weighted using techniques derived from multiple kernel learning. Because we treat the kernel weighting step as a differentiable convex optimization problem, our method can be regarded as an end-to-end learnable hybrid quantum-classical-convex neural network, or QCC-net, whose output is a data set-generalized kernel function suitable for use in any kernelized machine learning technique such as the support vector machine (SVM). Using our TSHK as input to a SVM, we classify univariate and multivariate time-series using quantum circuit simulators and demonstrate the efficient parallel deployment of the algorithm to 127-qubit superconducting quantum processors using quantum multi-programming.
△ Less
Submitted 17 February, 2024; v1 submitted 10 May, 2023;
originally announced May 2023.
-
A Quantum-Inspired Binary Optimization Algorithm for Representative Selection
Authors:
Anna G. Hughes,
Jack S. Baker,
Santosh Kumar Radha
Abstract:
Advancements in quantum computing are fuelling emerging applications across disciplines, including finance, where quantum and quantum-inspired algorithms can now make market predictions, detect fraud, and optimize portfolios. Expanding this toolbox, we propose the selector algorithm: a method for selecting the most representative subset of data from a larger dataset. The selected subset includes d…
▽ More
Advancements in quantum computing are fuelling emerging applications across disciplines, including finance, where quantum and quantum-inspired algorithms can now make market predictions, detect fraud, and optimize portfolios. Expanding this toolbox, we propose the selector algorithm: a method for selecting the most representative subset of data from a larger dataset. The selected subset includes data points that simultaneously meet the two requirements of being maximally close to neighboring data points and maximally far from more distant data points where the precise notion of distance is given by any kernel or generalized similarity function. The cost function encoding the above requirements naturally presents itself as a Quadratic Unconstrained Binary Optimization (QUBO) problem, which is well-suited for quantum optimization algorithms - including quantum annealing. While the selector algorithm has applications in multiple areas, it is particularly useful in finance, where it can be used to build a diversified portfolio from a more extensive selection of assets. After experimenting with synthetic datasets, we show two use cases for the selector algorithm with real data: (1) approximately reconstructing the NASDAQ 100 index using a subset of stocks, and (2) diversifying a portfolio of cryptocurrencies. In our analysis of use case (2), we compare the performance of two quantum annealers provided by D-Wave Systems.
△ Less
Submitted 4 January, 2023;
originally announced January 2023.
-
Origin of ferroelectric domain wall alignment with surface trenches in ultrathin films
Authors:
Jack S. Baker,
David R. Bowler
Abstract:
Engraving trenches on the surfaces of ultrathin ferroelectric (FE) films and superlattices promises control over the orientation and direction of FE domain walls (DWs). Through exploiting the phenomenon of DW-surface trench (ST) parallel alignment, systems where DWs are known for becoming electrical conductors could now become useful nanocircuits using only standard lithographical techniques. Desp…
▽ More
Engraving trenches on the surfaces of ultrathin ferroelectric (FE) films and superlattices promises control over the orientation and direction of FE domain walls (DWs). Through exploiting the phenomenon of DW-surface trench (ST) parallel alignment, systems where DWs are known for becoming electrical conductors could now become useful nanocircuits using only standard lithographical techniques. Despite this clear application, the microscopic mechanism responsible for the alignment phenomenon has remained elusive. Using ultrathin PbTiO$_3$ films as a model system, we explore this mechanism with large scale density functional theory simulations on as many as 5,136 atoms. Although we expect multiple contributing factors, we show that parallel DW-ST alignment can be well explained by this configuration giving rise to an arrangement of electric dipole moments which best restore polar continuity to the film. These moments preserve the polar texture of the pristine film, thus minimizing ST-induced depolarizing fields. Given the generality of this mechanism, we suggest that STs could be used to engineer other exotic polar textures in a variety of FE nanostructures as supported by the appearance of ST-induced polar cycloidal modulations in this letter. Our simulations also support experimental observations of ST-induced negative strains which have been suggested to play a role in the alignment mechanism.
△ Less
Submitted 29 October, 2021; v1 submitted 26 April, 2021;
originally announced April 2021.
-
A re-examination of antiferroelectric PbZrO$_3$ and PbHfO$_3$: an 80-atom $Pnam$ structure
Authors:
J. S. Baker,
M. Paściak,
J. K. Shenton,
P. Vales-Castro,
B. Xu,
J. Hlinka,
P. Márton,
R. G. Burkovsky,
G. Catalan,
A. M. Glazer,
D. R. Bowler
Abstract:
First principles density functional theory (DFT) simulations of antiferroelectric (AFE) PbZrO$_3$ and PbHfO$_3$ reveal a dynamical instability in the phonon spectra of their purported low temperature $Pbam$ ground states. This instability doubles the $c$-axis of $Pbam$ and condenses five new small amplitude phonon modes giving rise to an 80-atom $Pnam$ structure. Compared with $Pbam$, the stabilit…
▽ More
First principles density functional theory (DFT) simulations of antiferroelectric (AFE) PbZrO$_3$ and PbHfO$_3$ reveal a dynamical instability in the phonon spectra of their purported low temperature $Pbam$ ground states. This instability doubles the $c$-axis of $Pbam$ and condenses five new small amplitude phonon modes giving rise to an 80-atom $Pnam$ structure. Compared with $Pbam$, the stability of this structure is slightly enhanced and highly reproducible as demonstrated through using different DFT codes and different treatments of electronic exchange & correlation interactions. This suggests that $Pnam$ is a new candidate for the low temperature ground state of both materials. With this finding, we bring parity between the AFE archetypes and recent observations of a very similar AFE phase in doped or electrostatically engineered BiFeO$_3$.
△ Less
Submitted 21 February, 2021; v1 submitted 17 February, 2021;
originally announced February 2021.
-
Polar morphologies from first principles: PbTiO$_3$ films on SrTiO$_3$ substrates and the $p(2 \times Λ)$ surface reconstruction
Authors:
Jack S. Baker,
David R. Bowler
Abstract:
Low dimensional structures comprised of ferroelectric (FE) PbTiO$_3$ (PTO) and quantum paraelectric SrTiO$_3$ (STO) are hosts to complex polarization textures such as polar waves, flux-closure domains and polar skyrmion phases. Density functional theory (DFT) simulations can provide insight into this order, but, are limited by the computational effort needed to simulate the thousands of required a…
▽ More
Low dimensional structures comprised of ferroelectric (FE) PbTiO$_3$ (PTO) and quantum paraelectric SrTiO$_3$ (STO) are hosts to complex polarization textures such as polar waves, flux-closure domains and polar skyrmion phases. Density functional theory (DFT) simulations can provide insight into this order, but, are limited by the computational effort needed to simulate the thousands of required atoms. To relieve this issue, we use the novel multi-site support function (MSSF) method within DFT to reduce the solution time for the electronic groundstate whilst preserving high accuracy. Using MSSFs, we simulate thin PTO films on STO substrates with system sizes $>2000$ atoms. In the ultrathin limit, the polar wave texture with cylindrical chiral bubbles emerges as an intermediate phase between full flux closure domains and in-plane polarization. This is driven by an internal bias field born of the compositionally broken inversion symmetry in the [001] direction. Since the exact nature of this bias field depends sensitively on the film boundary conditions, this informs a new principle of design for manipulating chiral order on the nanoscale through the careful choice of substrate, surface termination or use of overlayers. Antiferrodistortive (AFD) order locally interacts with these polar textures giving rise to strong FE/AFD coupling at the PbO terminated surface driving a $p(2 \times Λ)$ surface reconstruction. This offers another pathway for the local control of ferroelectricity.
△ Less
Submitted 1 July, 2020;
originally announced July 2020.
-
The pseudoatomic orbital basis: electronic accuracy and soft-mode distortions in ABO$_3$ perovskites
Authors:
Jack S. Baker,
Tsuyoshi Miyazki,
David R. Bowler
Abstract:
The perovskite oxides are known to be susceptible to structural distortions over a long wavelength when compared to their parent cubic structures. From an ab initio simulation perspective, this requires accurate calculations including many thousands of atoms; a task well beyond the remit of traditional plane wave-based density functional theory (DFT). We suggest that this void can be filled using…
▽ More
The perovskite oxides are known to be susceptible to structural distortions over a long wavelength when compared to their parent cubic structures. From an ab initio simulation perspective, this requires accurate calculations including many thousands of atoms; a task well beyond the remit of traditional plane wave-based density functional theory (DFT). We suggest that this void can be filled using the methodology implemented in the large-scale DFT code, CONQUEST, using a local pseudoatomic orbital (PAO) basis. Whilst this basis has been tested before for some structural and energetic properties, none have treated the most fundamental quantity to the theory, the charge density $n(\mathbf{r})$ itself. An accurate description of $n(\mathbf{r})$ is vital to the perovskite oxides due to the crucial role played by short-range restoring forces (characterised by bond covalency) and long range coulomb forces as suggested by the soft-mode theory of Cochran and Anderson. We find that modestly sized basis sets of PAOs can reproduce the plane-wave charge density to a total integrated error of better than 0.5% and provide Bader partitioned ionic charges, volumes and average charge densities to similar degree of accuracy. Further, the multi-mode antiferroelectric distortion of PbZrO$_3$ and its associated energetics are reproduced by better than 99% when compared to plane-waves. This work suggests that electronic structure calculations using efficient and compact basis sets of pseudoatomic orbitals can achieve the same accuracy as high cutoff energy plane-wave calculations. When paired with the CONQUEST code, calculations with high electronic and structural accuracy can now be performed on many thousands of atoms, even on systems as delicate as the perovskite oxides.
△ Less
Submitted 11 March, 2020;
originally announced March 2020.
