Physics-Informed Bayesian Optimization of Variational Quantum Circuits
Authors: Kim Nicoli, Christopher J. Anders, Lena Funcke, Tobias Hartung, Karl Jansen, Stefan Kühn, Klaus-Robert Müller, Paolo Stornati, Pan Kessel, Shinichi Nakajima
NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our numerical experiments demonstrate that our approach improves over state-of-the-art baselines. and 4 Experiments We numerically demonstrate the performance of our approach for several setups... |
| Researcher Affiliation | Collaboration | Kim A. Nicoli1,2,3 Christopher J. Anders3,4 Lena Funcke1,2 Tobias Hartung5,6 Karl Jansen7 Stefan Kühn7 Klaus-Robert Müller3,4,8,9 Paolo Stornati10 Pan Kessel11 Shinichi Nakajima3,4,12 1Transdisciplinary Research Area (TRA) Matter, University of Bonn, Germany 2Helmholtz Institute for Radiation and Nuclear Physics (HISKP) 3Berlin Institute for the Foundations of Learning and Data (BIFOLD) 4 Technische Universität Berlin, Germany, 5 Northeastern University London, UK 6 Khoury College of Computer Sciences, Northeastern University, USA 7 CQTA, Deutsches Elektronen-Synchrotron (DESY), Zeuthen, Germany 8 Department of Artificial Intelligence, Korea University, Korea 9Max Planck Institut für Informatik, Saarbrücken, Germany 10Institute of Photonic Sciences, The Barcelona Institute of Science and Technology (ICFO) 11Prescient Design, g RED, Roche, Switzerland, 12RIKEN Center for AIP, Japan |
| Pseudocode | Yes | We refer the reader to Appendix F, in particular Algorithms 1 to 3, for the pseudo-codes and further algorithmic details complementing the brief introduction to the algorithms presented in the following sections. |
| Open Source Code | Yes | Our Python implementation along with detailed tutorials on how to reproduce the results is publicly available on Git Hub [44] at https://github.com/emicore/emicore. |
| Open Datasets | No | The paper uses classically simulated quantum computation for specific Hamiltonians (Ising and Heisenberg) and circuit configurations, which are defined within the paper, rather than relying on external, publicly available datasets. Therefore, no concrete access information for a public dataset is provided. |
| Dataset Splits | No | The paper describes training Gaussian Process (GP) models on sequentially observed data within the Bayesian Optimization framework, but it does not specify explicit training, validation, and test splits for a fixed dataset in the traditional machine learning sense or provide percentages/counts for such splits for its main VQE optimization experiments. |
| Hardware Specification | Yes | All numerical experiments have been performed on Intel Xeon Silver 4316 @ 2.30GHz CPUs |
| Software Dependencies | No | The paper mentions software like 'Qiskit' and 'SciPy' (e.g., 'The Qiskit [43] open-source library is used to classically simulate the quantum computer...' and 'In the code, the Sci Py [68] implementation of L-BFGS was used'), but it does not provide specific version numbers for these libraries or Python. |
| Experiment Setup | Yes | Appendix H and Table 4 provide specific hyperparameter values and training settings, including the number of readout shots, prior variance, smoothness parameter tuning schedules, and other parameters like TMI, JSG, JOG, NMC, TNFT, and TAve. |