Publications
The below list of publications should be up to date. Find me Google Scholar for more information.
P. A. M Casares, J. S. Baker, M. Medvidovic, R. D. Reis, J. M. Arrazola Grad DFT: a software library for machine learning enhanced density functional theory. arXiv preprint arXiv:2309.15127 (2023).
Summary: We introduce a Pythonic playground for experimentation with exchange & correlation functionals in DFT based on neural networks. Check out the code.
Summary: We introduce a Pythonic playground for experimentation with exchange & correlation functionals in DFT based on neural networks. Check out the code.
A. Ghukasyan, J. S. Baker, O. Goktas, J. Carrasquilla, S. K. Radha Quantum-Classical Multiple Kernel Learning. arXiv preprint arXiv:2305.17707 (2023).
Summary: We propose a nomenclature for classifying quantum-quantum, classical-classical and quantum-classical multiple kernel learning and conduct a series of numerical experiments. We introduce a new neural network architecture - quantum-classical-convex neural networks (QCC-nets) - which are able to train weighted combinations of parameterized classical and quantum kernels for any kernelized learning task.
Summary: We propose a nomenclature for classifying quantum-quantum, classical-classical and quantum-classical multiple kernel learning and conduct a series of numerical experiments. We introduce a new neural network architecture - quantum-classical-convex neural networks (QCC-nets) - which are able to train weighted combinations of parameterized classical and quantum kernels for any kernelized learning task.
J. S. Baker, G. Park, K. Yu, A. Ghukasyan, O. Goktas and, S. K. Radha Massively parallel hybrid quantum-classical machine learning for kernelized time-series classification. arXiv preprint arXiv:2305.05881 (2023).
Summary: A trainable time-dependent inner product space is devised using time evolution operators from quantum mechanics. We use this inner product space to define a time-series Hamiltonian kernel function, which, alongside classical SVM, we use to classify time-series on 127 qubit superconducting transmon chips. We speed up the algorithm by computing the kernel matrices in parallel using many quantum threads spread throughout the chip.
Summary: A trainable time-dependent inner product space is devised using time evolution operators from quantum mechanics. We use this inner product space to define a time-series Hamiltonian kernel function, which, alongside classical SVM, we use to classify time-series on 127 qubit superconducting transmon chips. We speed up the algorithm by computing the kernel matrices in parallel using many quantum threads spread throughout the chip.
K. Roleder, G. Catalan, A. M. Glazer, J. S. Baker, J. H. Ko, F. H. Naqvi, S. B. Junaid, A. Majchrowski, Z. Trybuła, J. Zaręba, I. Lazar, D. Kajewski, J. Koperski, and A. Soszyński, Weak low-temperature polarity in a PbZrO3 single crystal, Physical Review B 107, L140102 (2023).
Summary: We find experimental evidence for a new low temperature phase transition to a polar state in the purported archetypal antiferroelectric PbZrO3.
Summary: We find experimental evidence for a new low temperature phase transition to a polar state in the purported archetypal antiferroelectric PbZrO3.
H. Guo, J. S. Baker, W. Wu, and K. L. Choy, High Dielectric Constants in BaTiO3 Due to Phonon Mode Softening Induced by Lattice Strains: First Principles Calculations, Advanced Physics Research, 2300001 (2023).
Summary: Hybrid DFT caclulations show there exist certain strains where the dielectric constant of BaTiO3 is enhanced by a zone-centre phonon mode.
Summary: Hybrid DFT caclulations show there exist certain strains where the dielectric constant of BaTiO3 is enhanced by a zone-centre phonon mode.
A. G. Hughes, J. S. Baker, and S. K. Radha. A Quantum-Inspired Binary Optimization Algorithm for Representative Selection. arXiv preprint arXiv:2301.01836 (2023).
Summary: We develop a new binary objective function for representative selection and use it to perform portfolio diversification and index reconstruction using D-Wave quantum annealers.
Summary: We develop a new binary objective function for representative selection and use it to perform portfolio diversification and index reconstruction using D-Wave quantum annealers.
J. S. Baker, H. Horowitz, S. K. Radha, S. Fernandes, C. Jones, N. Noorani, V. Skavysh, P. Lamontangne, and B. C. Sanders. Quantum Variational Rewinding for Time Series Anomaly Detection. arXiv preprint arXiv:2210.16438 (2022).
Summary: A new quantum machine learning algorithm for time series anomaly detection is developed and used on superconducting transmon quantum hardware with advanced error mitigation.
Summary: A new quantum machine learning algorithm for time series anomaly detection is developed and used on superconducting transmon quantum hardware with advanced error mitigation.
J. S. Baker, and S. K. Radha. Wasserstein solution quality and the quantum approximate optimization algorithm: a portfolio optimization case study. arXiv preprint arXiv:2202.06782 (2022).
Summary: We explore a new way of looking at the quality of solutions from approximate optimizations algorithms like the QAOA and quantum annealing using Wasserstein distances. We use this new metric to benchmark the performance of several superconducting transmon and trapped ion quantum computers.
Summary: We explore a new way of looking at the quality of solutions from approximate optimizations algorithms like the QAOA and quantum annealing using Wasserstein distances. We use this new metric to benchmark the performance of several superconducting transmon and trapped ion quantum computers.
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 A re-examination of antiferroelectric PbZrO3 and PbHfO3: an 80-atom Pnam structure, arXiv preprint arXiv:2102.08856 (2021).
Summary: Surprisingly, the long-standing antiferroelectric groundstates of PbZrO3 and PbHfO3 are each found with an unstable phonon mode at 0K. The displacement pattern of this mode informs a new lower symmetry structure whoose energy is lower (according to DFT) than the presently supposed ground state.
Summary: Surprisingly, the long-standing antiferroelectric groundstates of PbZrO3 and PbHfO3 are each found with an unstable phonon mode at 0K. The displacement pattern of this mode informs a new lower symmetry structure whoose energy is lower (according to DFT) than the presently supposed ground state.
J. S. Baker and D. R. Bowler, Origin of ferroelectric domain wall alignment with surface trenches in ultrathin films, Physical Review Letters 127 247601 (2021).
Summary: We reveal the microscopic mechanism responsible for how ferroelectric domain walls align with surface trenches in ferroelectric films. We use large scale DFT calculations with thousands of atoms and thousands of physical cores using the UK national supercomputer: ARCHER2.
J. S. Baker and D. R. Bowler, Polar morphologies from first principles: PbTiO3 films on SrTiO3 substrates and the p(2 × Λ) surface reconstruction, Advanced Theory and Simulations 3(11) 2000154(2020).
Summary: Large scale DFT simulations are used to investigate polar vortices and polar waves in thin ferroelectric films. We find that asymmetrical polar textures can arise from broken inversion symmetries.
Summary: Large scale DFT simulations are used to investigate polar vortices and polar waves in thin ferroelectric films. We find that asymmetrical polar textures can arise from broken inversion symmetries.
A. Nakata, J. S. Baker, S. Y. Mujahed, J. TL Poulton, S. Arapan, J. Lin, Z. Raza, S.Yadav L. Truflandier, T. Miyazaki and D. R. Bowler, Large scale and linear scaling DFT with the CONQUEST code, The Journal of chemical physics 152(16) 164112 (2020).
Summary: A review of the large scale DFT code: CONQUEST. The theory behind the code is divulged and recent use cases are presented.
Summary: A review of the large scale DFT code: CONQUEST. The theory behind the code is divulged and recent use cases are presented.
J. S. Baker, T. Miyazaki and D. R. Bowler, The pseudoatomic orbital basis: electronic accuracy and soft-mode distortions in ABO3 perovskites, Electronic Structure, 2(2), 025002 (2020).
Summary: Using a selection of technologically relevant ferroelectric and antiferroelectric perovskite oxides, we study the suitability of DFT simulations using pseudoatomic orbital basis sets to describe their properties.
Summary: Using a selection of technologically relevant ferroelectric and antiferroelectric perovskite oxides, we study the suitability of DFT simulations using pseudoatomic orbital basis sets to describe their properties.
J. S. Baker, Long Range Order in Ferroelectric and Antiferroelectric Perovskites Meets Large Scale Density Functional Theory, University of London, University College London (United Kingdom), (2020).
Summary: My Ph.D. thesis. I studied long range behaviour in ferroelectric and antiferroelectric perovskite oxides using conventional and large scale DFT simulations. Systems of interest include bulk PZT, PbZrO3, PbHfO3 and thin films of PbTiO3. New emergent behviours were found including exotic polarization textures, new complex low energy structures in PbZrO3 and uncovering the mechanism behind the alignment of ferroelectric domain walls with engineered surface trenches.
Summary: My Ph.D. thesis. I studied long range behaviour in ferroelectric and antiferroelectric perovskite oxides using conventional and large scale DFT simulations. Systems of interest include bulk PZT, PbZrO3, PbHfO3 and thin films of PbTiO3. New emergent behviours were found including exotic polarization textures, new complex low energy structures in PbZrO3 and uncovering the mechanism behind the alignment of ferroelectric domain walls with engineered surface trenches.
J. S. Baker and D. R. Bowler, First-principles soft-mode lattice dynamics of PbZr0.5Ti0.5O3 and shortcomings of the virtual crystal approximation, Physical Review B 100, 224305 (2019)
Summary: New long wavelength crystal vibrations are found in the industrial piezoelectric PZT using DFT-based phonon calculations. We describe why the popular virtual crystal approximation cannot be used to accurately describe such patterns in PZT and similar materials.
Summary: New long wavelength crystal vibrations are found in the industrial piezoelectric PZT using DFT-based phonon calculations. We describe why the popular virtual crystal approximation cannot be used to accurately describe such patterns in PZT and similar materials.
D. R. Bowler, J. S. Baker, J. T. Poulton, S. Y. Mujahed, J. Lin, S. Yadav, Z. Raza and T. Miyazaki, Highly accurate local basis sets for large-scale DFT calculations in conquest, Japanese Journal of Applied Physics, 58(10), 100503 (2019).
Summary: We develop new heuristic methods for generating reliable basis sets of pseudoatomic orbitals for DFT calculations using CONQUEST. We study the accuracy of these basis sets using several popular materials.
Summary: We develop new heuristic methods for generating reliable basis sets of pseudoatomic orbitals for DFT calculations using CONQUEST. We study the accuracy of these basis sets using several popular materials.
L. K. McKemmish, K. L. Chubb, T. Rivlin, J. S. Baker, M. Gorman, A. Heward, W. Dunn and M.Tessenyi, Bringing pupils into the ORBYTS of research, Astronomy & Geophysics, 58(5), 5-11 (2017).
Summary: The Twinkle ORBYTS team discuss their approach to original research with school-aged scientists.
C. Sousa-Silva, L. K. McKemmish, K. L. Chubb, M. N. Gorman, J. S. Baker, E. J. Barton, T. Rivlin and J. Tennyson, Original Research By Young Twinkle Students (ORBYTS): when can students start performing original research?, Physics Education, 53(1), 015020 (2017).
Summary: The Twinkle ORBYTS team get involved in teaching research skills to school-aged scientists and have them publish articles in peer reviewed journals supervised by Ph.D. students and post-docs.
Summary: The Twinkle ORBYTS team get involved in teaching research skills to school-aged scientists and have them publish articles in peer reviewed journals supervised by Ph.D. students and post-docs.
