Erratum: Filter-function formalism and software package to compute quantum processes of gate sequences for classical non-Markovian noise [Phys. Rev. Research 3, 043047 (2021)]

Tobias Hangleiter, Pascal Cerfontaine, Hendrik BluhmPhysical 2024 · DOI · OpenAlex

Abstract

We claimed that our results for the noisy quantum process [Eqs.( 9) and ( 12)] are exact for Gaussian noise based on Kubo's paper on stochastic Liouville equations [1], in which he defines Gaussian operator processes by the property that only their first two cumulants are nonvanishing.However, as Fox pointed out 14 years later [2] and others have discussed since [3], Kubo's statement, founded on his paper on the cumulant expansion [4], does not hold for noncommuting stochastic variables such as, in general, the Hamiltonian in our case.The results presented in our paper are thus in general approximate up to order 2 .They become exact either for Hamiltonians whose deterministic part H c (t ) commutes with the stochastic part H n (t ) for all times t or for -correlated noise, that is, noise with a white power spectral density.Consequently, the following statements in the main text should be modified accordingly:(1) "We find exact results for the quantum process " in the abstract.(2) " which, for Gaussian noise, entails exact results for the quantum process " in Sec.I.(3) "In Sec.II we derive an exact expression for unital quantum operations " in Sec.I. ( 4) "If the noise is additionally Gaussian " at the beginning of Sec.II A 2.(5) "Remarkably, third-and higher-order cumulants vanish " before Eq. ( 12) and "Eqs.( 9) and ( 12) allow us to exactly compute " in Sec.II A 2.(6) "Note that Eq. ( 17) together with Eq. ( 9) constitutes an exact version " at the end of Sec.II A 2. ( 7) " we could have obtained the exact fidelity in the case of Gaussian noise." at the end of Sec.V A. ( 8) " we have shown how to compute the Liouville representation of the exact error channel " in Sec.VII. ( 9) "For Gaussian noise, our expansion is exact when including first-and second-order Magnus expansion (ME) terms.Hence, the convergence radius of the ME becomes infinite and the fidelity can be computed exactly by evaluating the matrix exponential Eq. ( 11).For non-Gaussian noise, the following considerations apply." in Appendix D.
BibTeX
@inproceedings{hangleiter2024erratum,
  author = {Tobias Hangleiter and Pascal Cerfontaine and Hendrik Bluhm},
  title = {{Erratum: Filter-function formalism and software package to compute quantum processes of gate sequences for classical non-Markovian noise [Phys. Rev. Research <b>3</b>, 043047 (2021)]}},
  booktitle = {Physical},
  year = {2024},
  doi = {10.1103/physrevresearch.6.049001},
}

Loading related content...

Figure