Binary Randomized Benchmarking (BRB)

Motivations

• Finding a protocol that is more scalable than CRB and DRB (removing starting and ending large circuits as in the DRB protocol).

Protocol

• The intial state \(\rho_0\) is obtained by is a single-gates layer that preprare the \(+1\) eigenstate of a uniformely randomly chosen \(n\)-qubit Pauli Operator.
• Each gate \(g_i\) is built from the user’s probability distribution \(\mu\) (as in DRB).
• The last layer is a single qubit gate layer that transform the quantum state into a tensor product of \(Z\) and \(I\) operators.
• The success metric is the probability of observing the \(+1\) eigenstate of the tensor product (the output is generally not deterministic, meaning that \(50\%\) of bitstrings are considered success and to others failure).

Limitations

• There is still scaling issues for large circuits as the final layer is computed from

Litterature: [1]

References

1. [1]J. Hines, D. Hothem, R. Blume-Kohout, B. Whaley, and T. Proctor, “Fully Scalable Randomized Benchmarking Without Motion Reversal,” PRX Quantum, vol. 5, no. 3, Aug. 2024, doi: 10.1103/prxquantum.5.030334. [Online]. Available at: http://dx.doi.org/10.1103/PRXQuantum.5.030334