arXiv:2610.31500
The Impact of Markovian Errors on Random Circuits
Anthony M. Polloreno, Timothy Proctor, Robin Blume-Kohout
Abstract
Randomized benchmarking (RB) returns a single decay parameter or gate error rate, but when the noise is gate-dependent that parameter is notoriously difficult to interpret. We study RB and related random-circuit benchmarks under small gate-local Markovian noise by factorizing each noisy gate into its ideal action followed by the exponential of a small error generator and then perturbing the RB $\mathscr{L}$ matrix, which models the RB decay. For any low-error gate set that satisfies the requirements of RB (it is a sequence-asymptotic unitary 2-design), we show that Pauli-stochastic error rates contribute linearly to the RB decay parameter, while Hamiltonian error rates contribute through a gauge-invariant quadratic form. Pauli-correlation and active generators are absent from the linear term and enter only at higher order. For the single-qubit native gate set $\{X_{π/2},Y_{π/2}\}$ we derive this quadratic form explicitly in terms of three first-order gauge-invariant (FOGI) coordinates: the two intrinsic over-rotations and one relational axis-tilt error. The resulting formula explains the dependence of DRB on the sampling distribution and the compilation dependence of CRB. These results turn RB decay rates into interpretable functions of experimentally meaningful elementary error rates.
AI Summary
Randomized benchmarking (RB) yields a scalar decay rate whose interpretation is unclear under gate-dependent noise. This paper asks how that rate depends on elementary coherent and dissipative errors, including the sampling distribution in direct RB (DRB) and compilation in Clifford RB (CRB). Each implemented gate is written as $e^{L_k}\mathcal G_k$, with its error generator decomposed into Hamiltonian, Pauli-stochastic, correlation, and active sectors. Perturbation theory for the RB $\mathscr L$ matrix gives, for any fixed sequence-asymptotic unitary 2-design, $r_e=\sum_{k,P}\Omega_k s_P^{(k)}+\frac{d^2-1}{d^2}Q_H(\mathbf h;\Omega)+O(\delta^2+\delta\theta+\theta^3)$, where $\delta$ and $\theta$ measure dissipative and Hamiltonian coefficients. Only Pauli-stochastic generators contribute linearly. The coherent quadratic form annihilates Hamiltonian gauge directions and therefore depends only on first-order gauge-invariant coordinates. For $\{X_{\pi/2},Y_{\pi/2}\}$, explicit DRB coefficients reduce six Hamiltonian parameters to two intrinsic over-rotations and one relational axis tilt. With equal over-rotations $\theta$ and uniform sampling, $r_e=3\theta^2/4+O(\theta^3)$, three times the leading native-gate infidelity. Pure tilt gives $r_e=p(1-p)\phi^2+O(\phi^3)$. An explicit Clifford-level quadratic form, combined with first-order propagation through compiled sequences, explains compilation dependence; exact transfer-matrix calculations check the approximations. The assumptions exclude drift, temporal correlations, context dependence, and unmodeled leakage. The perturbative neighborhood depends on the sampling distribution and is not uniform as its design gap closes. Explicit formulas cover one-qubit examples; larger gate sets and higher-order correlation and active contributions remain open. Multiple RB rates constrain, but do not uniquely reconstruct, signed coherent errors.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 15 pages, 3 figures