Misty Wahl

Removed #1156 from 0.15.0 milestone - solve #1214 in 0.15.0 milestone instead (incremental fix).

@crazy4pi314 I'd like to give this one a try.

Link to my draft RFC - still very rough https://docs.google.com/document/d/1Bl8Ix4MB32wa_X_FwMx2LLmt3BRyzoyHwcuYphbGIV8/edit?usp=sharing

Draft RFC https://docs.google.com/document/d/13kD3O-sMZARgSTNItTn58hzWbDLs3I0WtFL3nZffh-M/edit

https://docs.google.com/document/d/1VrZRV59-2ApemebfbL556Rj7I2AMhAS7bJgGuTg3K1w/edit?usp=sharing

Step 1 of implementation complete: function to calculate biased noise representations (see #1109)

As demonstrated in PR #1364, multiple frontends are supported and therefore the learning and loss functions are made public.

> * 1. Seed in tests may help with avoid stochastic fails in continuous integration and pin down the problem > * 2. Many files are changed. Rebasing against master...

@andreamari @nathanshammah, some good news: tests for Rx, Rz, and CNOT, for `epsilon` 0.05 and 0.1 are passing! The bad news is the execution time is still too long, especially...

> @Misty-W > > > 2. Find ways of shortening execution time --> suggestions welcome! > > Have you tried setting initial conditions for epsilon closer or (even equal) to...