Dataset for Quantum Circuits Mapping

Published: 16 September 2021| Version 1 | DOI: 10.17632/pmycgb2bt7.1
Contributors:
Roberto Schiattarella,

Description

The proposed dataset helps to address the quantum circuit mapping problem as a classification task. Three csv files are provided, each one refers to a specific IBM quantum machine, namely IBMQ Santiago, IBMQ Athens, and IBMQ 16 Melbourne. Each csv file is composed of random quantum circuits mapped onto the specific IBM quantum processor. In detail, each dataset instance contains some features related to the calibration data of the physical device and others related to the generated quantum circuit. Finally, the instance is labeled with a vector encoding the best mapping among those provided by deterministic mapping algorithms.

Files

Steps to reproduce

The collection of data can be divided into four steps: 1) Circuit Generation and DateSelection: a random quantum circuit is generated by Qiskit and simultaneously a date is selected; 2) Circuit Features Extraction: a set of information related to the generated quantum circuit is extracted;3) Processor Features Extraction: the calibration data provided by IBM for the date selected in step 1, are extracted;4) Label Selection: the best deterministic mapping provided by qiskit transpiler is selected as the label of the instance.

Categories

Quantum Computing, Machine Learning

Licence