Randomized benchmarking with online adaptive heuristic

This sample demonstrates using Q# to run a randomized benchmarking experiment with an online heuristic for choosing sequence lengths. In particular, this sample uses a statistical inference algorithm written in Q# to choose randomized benchmarking sequences so as to maximize how much we learn about the fidelity of a gateset with each sequence.

NOTE: This sample is somewhat more advanced than most, and assumes familiarity with open quantum systems as well as some statistics knowledge. If you would like a refresher on these topics, please check out the process-tomography and phase-estimation samples.

Installation

As this sample demonstrates using Q# and Python together, make sure you have the qsharp package for Python installed first; see the Getting Started with Python guide for details.

This sample also uses a couple extra Python packages to help out, so you'll need to have those ready as well. If you are using the Anaconda distribution of Python, this can be done automatically by using the environment.yml file provided with this sample:

cd samples/characterization/randomized-benchmarking
conda env create -f environment.yml
conda activate randomized-benchmarking

Running the Sample

Once everything is installed, run jupyter notebook to start the Jupyter Notebook interface in your web browser:

jupyter notebook

In the browser, select the randomized-benchmarking.ipynb notebook in your browser to view the sample.

Manifest

• randomized-benchmarking.ipynb: Jupyter Notebook demoing the randomized benchmarking protocol.
• environment.yml: Specification of a conda environment for use with randomized benchmarking.
• Math.qs: Special functions and linear algebra support for this sample.
• Inference.qs: Implementation of the particle filtering algorithm for Q#.