Add an example in which a two-qubit gate error is empirically estimated for implementing PEC

[andreamari]

Issue Description

PEC requires the knowledge of the noise model, or at least a simplified approximation of it. A crude approximation is assuming that noise is dominant for two-qubit gates and that it can be approximated by a depolarizing channel. Even this crude approximation, still requires knowing at least the size of a real parameter: the two-qubit gate error.

This parameter could be deduced from the hardware properties as suggested in #1400 . Alternatively, it would be nice to show a more empirical method in which one can estimate such two-qubit error probability.

Proposed Solution

A brute force method could be performing a gate tomography. However, it would be nice to show a simpler empirical way.

E.g. Evaluating the survival probability for circuits like (CNOT CNOT)^n, fitting an exponential decay, deducing the two-qubit gate error from the exponent of the exponential decay.

More advanced or more rigorous ways are welcome, as long as they are not too involved to implement in practice.