Researchers introduce surrogate-enabled zero-noise extrapolation (S-ZNE), a quantum error mitigation technique that uses classical learning surrogates to perform zero-noise extrapolation on the classical side rather than through repeated quantum measurements. Unlike conventional zero-noise extrapolation, whose measurement cost scales linearly with the number of circuits in a parameterized family, S-ZNE needs only constant measurement overhead regardless of how many circuits are involved. Theoretical analysis and numerical experiments on ground-state energy and quantum metrology tasks with up to 100 qubits show accuracy comparable to conventional zero-noise extrapolation, suggesting the approach could generalize to other quantum error mitigation protocols.
Questions this post answers
What is surrogate-enabled zero-noise extrapolation (S-ZNE) in quantum error mitigation?
S-ZNE is a quantum error mitigation technique that performs zero-noise extrapolation entirely on the classical side using classical learning surrogates instead of repeated quantum measurements. While conventional zero-noise extrapolation has measurement cost scaling linearly with the number of circuits in a parameterized family, S-ZNE requires only constant measurement overhead for the entire family, and numerical experiments on up to 100-qubit ground-state energy and quantum metrology tasks confirm accuracy comparable to conventional zero-noise extrapolation. Researchers tracking advances in quantum error mitigation can follow developments like this on daily.dev.