Supervised Learning Enhanced Quantum Circuit Transformation

TL;DR

This paper introduces a supervised learning framework using neural networks to enhance quantum circuit transformation, improving efficiency and performance in adapting circuits to quantum hardware constraints.

Contribution

It presents a novel approach that integrates trained neural networks into existing QCT algorithms, enabling more effective SWAP gate selection without significant additional computational cost.

Findings

Improved transformation performance across various QPU connectivities.

Neural networks trained on shallow circuits generalize well to complex circuits.

Enhanced fidelity and reduced circuit depth in transformed quantum circuits.

Abstract

A quantum circuit transformation (QCT) is required when executing a quantum program in a real quantum processing unit (QPU). Through inserting auxiliary SWAP gates, a QCT algorithm transforms a quantum circuit to one that satisfies the connectivity constraint imposed by the QPU. Due to the non-negligible gate error and the limited qubit coherence time of the QPU, QCT algorithms which minimize gate number or circuit depth or maximize the fidelity of output circuits are in urgent need. Unfortunately, finding optimized transformations often involves exhaustive searches, which are extremely time-consuming and not practical for most circuits. In this paper, we propose a framework that uses a policy artificial neural network (ANN) trained by supervised learning on shallow circuits to help existing QCT algorithms select the most promising SWAP gate. ANNs can be trained off-line in a distributed way. The trained ANN can be easily incorporated into QCT algorithms without bringing too much overhead in time complexity. Exemplary embeddings of the trained ANNs into target QCT algorithms demonstrate that the transformation performance can be consistently improved on QPUs with various connectivity structures and random or realistic quantum circuits.