Error mitigation service for NISQ devices
The error mitigation service can mitigate the impact of errors of noisy measurement results obtained from a quantum computer.
The service enables the generation and management of calibration and mitigation data for multiple QPU providers. Further, it allows users to mitigate their execution results on the basis of freshly generated or already existing mitigation data, speeding up the typical error mitigation process. The error mitigation service currently implements multiple methods, such as Mthree or TPNM for IBMQ and IonQ. Additional mitigation methods and QPU providers can easily be integrated into the service.
Running the Application
Run the App with Docker-Compose
docker-compose up
API Structure
In the following the structure of the error mitigation service API is explained briefly. The full API specification can be found here when the service is running.
/cm
Endpoint for generating calibration matrices by generating and running calibration circuits on QPUs.
/mm
Endpoint for generating mitigators based on calibration data. These mitigators can then be used to mitigate execution results.
Endpoint for generating mitigators based on calibration data. If no calibration data is available, it will automatically be generated if the required QPU credentials are included in the request. Mitigators can be used to mitigate execution results.
/rem
Endpoint for mitigating results retrieved from circuit executions on NISQ devices.
Endpoint for mitigating results retrieved from circuit executions on NISQ devices. If no calibration data is available, it will automatically be generated if the required QPU credentials are included in the request.
Developer Guide
Setup (exemplary for ubuntu 18.04):
git clone https://github.com/UST-QuAntiL/error-mitigation-service.gitcd error-mitigation-service-
sudo -H pip install virtualenv(if you don't have virtualenv installed) -
virtualenv venv(create virtualenv named 'venv') -
source venv/bin/activate(enter virtualenv; in Windows systems activate might be invenv/Scripts) -
pip install -r requirements.txt(install application requirements)