This package is intended to be used for studying the response of different quantum error correction schemes on various types of quantum noise processes. Here, numerical simulations can be peformed for any quantum (stabilizer) error correcting scheme over any noise model, thereby revealing how different measures of noise strength for a logical qubit depend on physical noise parameters.

As an example, the following plot produced by chflow describes the fidelity of a logical qubit error corrected for coherent errors (rotations about Z-axis) as a function of the angle of the over rotation error.

The following command set is required to generate the above plot.

Pavithrans-MacBook-Pro:chflow pavithran$ ./chflow.sh
>> sbload 11_04_2018_13_52_37
Simulation data is available for 0% of the channels.
Preparing physical channels... 26 (100%) done.        
 >> ecc
Simulation data is available for 0% of the channels.
Please wait ...
100% done, approximately 0 seconds remaining ...   
done, in 870 seconds.
>> metrics fidelity 
>> collect
Simulation data is available for 100% of the channels.
>> lplot 0 fidelity
>> 

For any stabilizer error correction scheme, chflow tracks the flow of a quantum channel -- evolution from a physical noise process to the effective quantum channel that affects the underlying logical information encompassing quantum error correction steps. To learn more about such techniques, see arXiv:1711.04736 and Phys. Rev. A 95, 042332.

In addition to its main purpose, this package can also be used to perform simple but useful operations on quantum error correcting codes as well as quantum channels. For instance, derivation of the canonical basis (Stabilizers, Logicals and Pure errors) for an error correcting code, converting between various representations of a quantum channel, computing Pauli approximations of a quantum channel and so on.

The latest version of chflow can be obtained by cloning this repository. All of the interface and plotting tools require Python (any version higher than) 2.7 and standard packages such as Numpy, Scipy and Matplotlib. On the other hand, most of the error correction simulations that run in the backend are implemented as Cython extensions, with the associated C files already provided. However, they can also be generated using the build cython command in chflow, which requires Cython and gcc compilers. Some advanced functions require the PICOS, CVXOPT and scikit-learn packages. When chflow is started, a text file called requirements.txt will be generated in chflow/ containing the names of the missing packages. To fulfill all of those requirements, one can simply run pip install -r requirements.txt on the shell. MacOS users must prepend this with a sudo. Refer to the Wiki for instructions on running chflow.

This material is based upon work supported by, or in part by, the U.S. Army Research Laboratory and the U.S. Army Research Office under Contract W911NF-14-C-0048.

There are no restrictions on contributing to chflow. While updating the GitHub repository, care should be taken to avoid syncing data files and compiler generated files. Two commands, clean and clean git, in chflow remove such unwanted files. The latter option returns the chflow/ directory to its factory setting. Note that when files are removed, they are not actually deleted, but moved into chflow/.ignore. It is then safe to delete this folder using rm -r .ignore/.

P. Iyer and D. Poulin. chflow: quantum error correction for realistic noise. https://github.com/paviudes/chflow, DOI: 10.5281/zenodo.1216202, April 2018.

@misc{chflow,
  author       = {Pavithran Iyer, Aditya Jain and David Poulin},
  title        = {chflow: Quantum error correction for realistic noise},
  month        = {April},
  year         = {2023},
  doi          = {10.5281/zenodo.1216202},
  url          = {https://github.com/paviudes/chflow}
}

Comments and suggestions are welcome, they can be addressed to [email protected].