WFW. Matthew C. Foulkes

W. Matthew C. Foulkes

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


W. Matthew C. Foulkes has a total of 88 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Extracting Anyon Statistics from Neural Network Fractional Quantum Hall States

    by Andres Perez Fadon, David Pfau, James S. Spencer, Wan Tong Lou, Titus Neupert, W. M. C. Foulkes

    Fractional quantum Hall states host emergent anyons with exotic exchange statistics, but obtaining direct access to their topological properties in real systems remains a challenge. Neural network wave functions provide a flexible computational approach, as they can represent highly correlated states without requiring a tailored basis. Here, we use the neural network variational Monte Carlo method to study the fractional quantum Hall effect on the torus and find the three degenerate ground states at filling factor ν = 1 / 3 . From these, we extract the modular S -matrix via entanglement interferometry, a technique previously only applied to lattice models. The resulting S -matrix encodes the quantum dimensions, fusion rules, and exchange statistics of the emergent anyons, providing a direct numerical demonstration of the topological order. The calculated anyon properties match the well-known theoretical and experimental results. Our work establishes neural network wave functions as a powerful tool for investigating anyonic properties.


  • Open Access

    Neural wavefunction calculations of hyperfine coupling constants relevant to μSR spectroscopy with quantum muons and protons

    by Jamie Carr, Mathias Volkai, W. M. C. Foulkes, Andres Perez Fadon

    Accurate prediction of muon hyperfine constants at zero temperature is useful for interpreting muon spin spectroscopy data, yet standard methods, such as density functional theory (DFT), evaluate these constants within the Born–Oppenheimer approximation, in which the electronic wavefunction is calculated with the muon and nuclei held as clamped point charges. This treatment may be inadequate for the muon because of its small mass. To overcome this limitation, we use the variational quantum Monte Carlo method with neural-network trial wavefunctions, a highly accurate and flexible approach recently applied to other quantum chemical problems. Within this framework, the muon can either be clamped or included explicitly in the many-particle electron-muon wavefunction, in which case the fully quantum-mechanical pair density is obtained directly. We calculate zero-temperature vacuum muon hyperfine constants in muoniated methyl and ethyl radicals for both clamped and quantum-mechanical muons. The hyperfine constants obtained from our clamped-muon calculations in the methyl and ethyl radicals differ significantly from the corresponding DFT results, highlighting the limitations of DFT even when the muon is clamped. The results with quantum muons are closer to experiment after accounting for environmental and temperature effects. These findings suggest that explicitly calculating the quantum-mechanical muon-electron pair density improves the accuracy of muon hyperfine constant predictions.

  • Open Access

    Transferable neural wavefunctions for solids

    by L. Gerard, M. Scherbela, H. Sutterud, W. M. C. Foulkes, P. Grohs

    Deep-learning-based variational Monte Carlo has emerged as a highly accurate method for solving the many-electron Schrödinger equation. Despite favorable scaling with the number of electrons, $${\mathcal{O}}({{n}_{{\rm{el}}}}^{4})$$ O ( n el 4 ) , the practical value of deep-learning-based variational Monte Carlo is limited by the high cost of optimizing the neural network weights for every system studied. Recent research has proposed optimizing a single neural network across multiple systems, reducing the cost per system. Here we extend this approach to solids, which require numerous calculations across different geometries, boundary conditions and supercell sizes. We demonstrate that optimization of a single ansatz across these variations significantly reduces optimization steps. Furthermore, we successfully transfer a network trained on 2 × 2 × 2 supercells of LiH, to 3 × 3 × 3 supercells, reducing the number of optimization steps required to simulate the large system by a factor of 50 compared with previous work.


  • Interaction-induced symmetry breaking in circular quantum dots

    by Andres Perez Fadon, Gino Cassella, Halvard Sutterud, W. M. C. Foulkes

    This paper investigates interaction-induced symmetry breaking in circular quantum dots. We start by explaining what is known about symmetry breaking in quantum dots, pointing out that the anisotropic “static Wigner molecule” ground states frequently observed in simulations are created by interference effects that occur even in the non-interacting limit. They have nothing in common with the interaction-driven crystallization of the uniform electron gas described by Wigner. This leads us to define the term Wigner molecule more carefully via a finite analog of the spontaneous symmetry breaking that arises in the homogeneous electron gas when the interactions are strong. According to this definition, the charge density patterns characteristic of true interaction-induced Wigner molecules can only be seen if a small symmetry-breaking perturbation is applied to a strongly interacting quantum dot. A simple argument based on separation of variables into center-of-mass and internal coordinates shows that the strength of the perturbation required to produce a finite effect on the density tends to zero in the limit as the strength of the interaction tends to infinity. We confirm computationally that interaction-induced Wigner molecules satisfying this definition exist. The neural-network variational Monte Carlo method used in our simulations proves more accurate than the coupled cluster and diffusion Monte Carlo methods employed in previous benchmark calculations of quantum dots at small to intermediate interaction strengths. For high interaction strengths, our neural-network variational Monte Carlo energies agree very well with existing fixed-node diffusion Monte Carlo benchmarks, proving ∼0.01% better for small values of the total spin projection Sz but ∼0.01% worse for fully spin-polarized systems.


  • Open Access

    Neural network variational Monte Carlo for positronic chemistry

    by Gino Cassella, W. M. C. Foulkes, David Pfau, James S. Spencer

    Quantum chemical calculations of the ground-state properties of positron-molecule complexes are challenging. The main difficulty lies in employing an appropriate basis set for representing the coalescence between electrons and a positron. Here, we tackle this problem with the recently developed Fermionic neural network (FermiNet) wavefunction, which does not depend on a basis set. We find that FermiNet produces highly accurate, in some cases state-of-the-art, ground-state energies across a range of atoms and small molecules with a wide variety of qualitatively distinct positron binding characteristics. We calculate the binding energy of the challenging non-polar benzene molecule, finding good agreement with the experimental value, and obtain annihilation rates which compare favourably with those obtained with explicitly correlated Gaussian wavefunctions. Our results demonstrate a generic advantage of neural network wavefunction-based methods and broaden their applicability to systems beyond the standard molecular Hamiltonian.


  • Open Access

    Neural Wave Functions for Superfluids

    by Wan Tong Lou, Halvard Sutterud, Gino Cassella, W. M. C. Foulkes, Johannes Knolle, David Pfau, James S. Spencer

    Understanding superfluidity remains a major goal of condensed matter physics. Here, we tackle this challenge utilizing the recently developed fermionic neural network (FermiNet) wave function Ansatz [D. Pfau et al., Phys. Rev. Res. 2, 033429 (2020).] for variational Monte Carlo calculations. We study the unitary Fermi gas, a system with strong, short-range, two-body interactions known to possess a superfluid ground state but difficult to describe quantitatively. We demonstrate key limitations of the FermiNet Ansatz in studying the unitary Fermi gas and propose a simple modification based on the idea of an antisymmetric geminal power singlet (AGPs) wave function. The new AGPs FermiNet outperforms the original FermiNet significantly in paired systems, giving results which are more accurate than fixed-node diffusion Monte Carlo and are consistent with experiment. We prove mathematically that the new Ansatz, which differs from the original Ansatz only by the method of antisymmetrization, is a strict generalization of the original FermiNet architecture, despite the use of fewer parameters. Our approach shares several advantages with the original FermiNet: The use of a neural network removes the need for an underlying basis set; sand the flexibility of the network yields extremely accurate results within a variational quantum Monte Carlo framework that provides access to unbiased estimates of arbitrary ground-state expectation values. We discuss how the method can be extended to study other superfluid.


  • Open Access

    The Einstein-de Haas effect in an Fe15 cluster

    by T Wells, W M C Foulkes, S L Dudarev, A P Horsfield

    Classical models of spin-lattice coupling are at present unable to accurately reproduce results for numerous properties of ferromagnetic materials, such as heat transport coefficients or the sudden collapse of the magnetic moment in hcp-Fe under pressure. This inability has been attributed to the absence of a proper treatment of effects that are inherently quantum mechanical in nature, notably spin–orbit coupling (SOC). This paper introduces a time-dependent, non-collinear tight binding model, complete with SOC and vector Stoner exchange terms, that is capable of simulating the Einstein-de Haas (EdH) effect in a ferromagnetic Fe15 cluster. The tight binding model is used to investigate the adiabaticity timescales that determine the response of the orbital and spin angular momenta to a rotating, externally applied B field, and we show that the qualitative behaviors of our simulations can be extrapolated to realistic timescales by use of the adiabatic theorem. An analysis of the trends in the torque contributions with respect to the field strength demonstrates that SOC is necessary to observe a transfer of angular momentum from the electrons to the nuclei at experimentally realistic B fields. The simulations presented in this paper demonstrate the EdH effect from first principles using a Fe cluster.

  • Top Cited Papers

    Ab initio quantum chemistry with neural-network wavefunctions

    by Jan Hermann, James Spencer, Kenny Choo, Antonio Mezzacapo, W. M. C. Foulkes, David Pfau, Giuseppe Carleo, Frank Noé

    Deep learning methods outperform human capabilities in pattern recognition and data processing problems and now have an increasingly important role in scientific discovery. A key application of machine learning in molecular science is to learn potential energy surfaces or force fields from ab initio solutions of the electronic Schrödinger equation using data sets obtained with density functional theory, coupled cluster or other quantum chemistry (QC) methods. In this Review, we discuss a complementary approach using machine learning to aid the direct solution of QC problems from first principles. Specifically, we focus on quantum Monte Carlo methods that use neural-network ansatzes to solve the electronic Schrödinger equation, in first and second quantization, computing ground and excited states and generalizing over multiple nuclear configurations. Although still at their infancy, these methods can already generate virtually exact solutions of the electronic Schrödinger equation for small systems and rival advanced conventional QC methods for systems with up to a few dozen electrons.


  • Top Cited Papers
    Open Access

    Discovering Quantum Phase Transitions with Fermionic Neural Networks

    by Gino Cassella, Halvard Sutterud, Sam Azadi, N. D. Drummond, David Pfau, James S. Spencer, W. M. C. Foulkes

    Deep neural networks have been very successful as highly accurate wave function Ansätze for variational Monte Carlo calculations of molecular ground states. We present an extension of one such Ansatz, FermiNet, to calculations of the ground states of periodic Hamiltonians, and study the homogeneous electron gas. FermiNet calculations of the ground-state energies of small electron gas systems are in excellent agreement with previous initiator full configuration interaction quantum Monte Carlo and diffusion Monte Carlo calculations. We investigate the spin-polarized homogeneous electron gas and demonstrate that the same neural network architecture is capable of accurately representing both the delocalized Fermi liquid state and the localized Wigner crystal state. The network converges on the translationally invariant ground state at high density and spontaneously breaks the symmetry to produce the crystalline ground state at low density, despite being given no a priori knowledge that a phase transition exists.


  • Electronic structure of α−Al2O3 grain boundaries containing reactive element segregants

    by Andy Paul Chen, Arthur H. Heuer, Michael W. Finnis, W. M. C. Foulkes

    It has long been known that the addition of small quantities (“doping”) of so-called reactive elements (REs) such as Y, Zr, and Hf to high-temperature Al2O3 scale-forming alloys improves oxidation resistance. The presence of reactive elements at grain boundaries lowers the growth rate of the α−Al2O3 scales, but the cause of the reduced scale growth kinetics is not fully understood. Explanations based on steric effects and explanations based on reducing the grain boundary electronic conductivity have been proposed. We have used density functional theory to study the structural and electronic properties of two Σ7 bicrystal grain boundaries containing Y, Hf, and Zr substitutional defects on Al sites. The presence of RE substitutional defects plays a minimal direct role in reducing the density of electronic states near the valence-band maximum. However, Hf4+ or Zr4+ substitutions at the grain boundary repel the positively charged oxygen vacancy VO2+. As VO2+ contributes a defect state above the valence-band maximum but below the Fermi energy, this indirectly lowers the density of current-carrying holes and thus the electronic conductivity of the grain boundary. Replacing Al3+ ions with Hf4+ or Zr4+ ions also makes the grain boundary positively charged, further reducing the hole density.