Error-mitigated quantum computation with noisy superconducting qubits*

In the absence of fault tolerant hardware architectures, the development of hybrid algorithms has led to interest in approximate computing with noisy intermediate scale quantum computers. However, recent demonstrations [1] have been severely limited by decoherence, highlighting the need for error mitigation techniques to extract accurate computations from noisy quantum hardware. In this context, I shall introduce a zero-noise extrapolation technique [2,3] that enables access to noise-free estimates of expectation values, after the application of a short depth circuit, without requiring any additional quantum resources. I shall discuss challenges for its implementation with superconducting qubits [4], and highlight its broad applicability with experimental demonstrations of quantum simulation [4] and machine learning [5]. [1] A. Kandala, A. Mezzacapo et al, Nature 549, 242 (2017) [2] K. Temme et al, PRL 119, 180509 (2017) [3] Li et al, PRX 7, 021050 (2017) [4] A. Kandala et al, arXiv:1805.04492 [5] V. Havelick et al, arXiv:1804.11326