Error-mitigated data-driven circuit learning on noisy quantum hardware
- PDF / 3,783,005 Bytes
- 15 Pages / 595.224 x 790.955 pts Page_size
- 4 Downloads / 221 Views
RESEARCH ARTICLE
Error-mitigated data-driven circuit learning on noisy quantum hardware Kathleen E. Hamilton1
· Raphael C. Pooser1
Received: 11 December 2019 / Accepted: 19 June 2020 © Springer Nature Switzerland AG 2020
Abstract Application-level benchmarks measure how well a quantum device performs meaningful calculations. In the case of parameterized circuit training, the computational task is the preparation of a target quantum state via optimization over a loss landscape. This is complicated by various sources of noise, fixed hardware connectivity, and generative modeling, the choice of target distribution. Gradient-based training has become a useful benchmarking task for noisy intermediate-scale quantum computers because of the additional requirement that the optimization step uses the quantum device to estimate the loss function gradient. In this work, we use gradient-based data-driven circuit learning to qualitatively evaluate the performance of several superconducting platform devices and present results that show how error mitigation can improve the training of quantum circuit Born machines with 28 tunable parameters. Keywords Quantum machine learning · Error mitigation
1 Introduction The field of quantum machine learning covers a diverse range of topics, from utilizing quantum computing to speed up classical model training (Harrow et al. 2009; Potok et al. 2018) to using quantum circuits as analogues of classical models. Parameterized quantum circuits have become a popular approach to construct general trainable quantum models. These circuits can find use in state preparation for algorithms compatible with noisy intermediate-scale quantum (NISQ) hardware. The training process for these circuits is similar to the variational quantum eigensolver, This manuscript has been authored by UT-Battelle, LLC, under Contract No. DE-AC0500OR22725 with the US Department of Energy. The US Government retains and the publisher, by accepting the article for publication, acknowledges that the US Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for the US Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan. Kathleen E. Hamilton
[email protected] 1
Oak Ridge National Laboratory, One Bethel Valley Road, Oak Ridge, TN, 37830 USA
and it can be supervised, which has been used for classifiers (Mitarai et al. 2018; Havl´ıcˇ ek et al. 2019), unsupervised, and it can be used for generative models (Benedetti et al. 2019; Liu and Wang 2018b). To extract accurate performance from NISQ devices, a wide range of techniques and methods have been developed which fall under the term “error mitigation” (EM). In the absence of noise, circuits with few qubits and entangling layers are able to model general discrete and continuous distributions. However, realizing these modeling properties on current NISQ
Data Loading...
