Currently, only discrimination is actually supported, either averaged or through shots.

However, this can be upgraded to integration just placing the $\ket{1}$ in the $(1, 0)$ point in the IQ plane, and $\ket{0}$ in the $(0, 0)$ point, artificially.

Simulate improvements

Minor improvements may be possible just configuring some values related to the readout emulation, without actually performing any emulation at all.

• we could improve shots just by summing some Gaussian noise, whose width should be configured somewhere
• we could rotate and scale the points in the plane, reading scaling factor and rotation angles from some configurations