pyGSTio/pyGSTi: Version 0.9.9.3

Compressive gate set tomography

“…To this end, we construct the measurement sequences from a number of N f fiducial sequences of length 3 drawn uniformly at random. We base our gate set on the so-called XYI model, the standard single qubit example in the pyGSTi package [38]. The XYI model consists of the identity gate, a π/2 X-rotation and a π/2 Y-rotation on the Bloch sphere, with initial state |0 0 | and measurement in the computational basis.…”

“…In numerical simulations we demonstrate that the structure-exploiting mGST (i) allows for maximal flexibility in the design of gate sequences, so that standard GST gate sequences and random sequences work equally well, and (ii) obtains lowrank approximations of the implemented gate set from a significantly reduced number of sequences and samples. This allows us to successfully perform GST with gate sets and sequences that are not amenable to the standard GST implementation pyGSTi [15,38]. As one example, while the sequence design of pyGSTi uses at least 907 specific sequences to reconstruct a two-qubit gate set, we numerically demonstrate low-rank reconstruction from 200 random sequences of maximal length 7 with runtimes of less than an hour on a standard desktop computer.…”