Provable bounds for noise-free expectation values computed from noisy samples.

Autor: Barron SV; IBM Quantum, IBM Thomas J. Watson Research Center, Yorktown Heights, NY, USA., Egger DJ; IBM Quantum, IBM Research Europe-Zurich, Rueschlikon, Switzerland., Pelofske E; CCS-3 Information Sciences, Los Alamos National Laboratory, Los Alamos, NM, USA.; A-1 Information Systems & Modeling, Los Alamos National Laboratory, Los Alamos, NM, USA., Bärtschi A; CCS-3 Information Sciences, Los Alamos National Laboratory, Los Alamos, NM, USA., Eidenbenz S; CCS-3 Information Sciences, Los Alamos National Laboratory, Los Alamos, NM, USA., Lehmkuehler M; University of Basel, Basel, Switzerland., Woerner S; IBM Quantum, IBM Research Europe-Zurich, Rueschlikon, Switzerland. wor@zurich.ibm.com.
Jazyk: angličtina
Zdroj: Nature computational science [Nat Comput Sci] 2024 Nov; Vol. 4 (11), pp. 865-875. Date of Electronic Publication: 2024 Nov 01.
DOI: 10.1038/s43588-024-00709-1
Abstrakt: Quantum computing has emerged as a powerful computational paradigm capable of solving problems beyond the reach of classical computers. However, today's quantum computers are noisy, posing challenges to obtaining accurate results. Here, we explore the impact of noise on quantum computing, focusing on the challenges in sampling bit strings from noisy quantum computers and the implications for optimization and machine learning. We formally quantify the sampling overhead to extract good samples from noisy quantum computers and relate it to the layer fidelity, a metric to determine the performance of noisy quantum processors. Further, we show how this allows us to use the conditional value at risk of noisy samples to determine provable bounds on noise-free expectation values. We discuss how to leverage these bounds for different algorithms and demonstrate our findings through experiments on real quantum computers involving up to 127 qubits. The results show strong alignment with theoretical predictions.
Competing Interests: Competing interests: The authors declare no competing interests.
(© 2024. IBM.)
Databáze: MEDLINE
Externí odkaz: