Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Predicting Good Quantum Circuit Compilation Options

Nils Quetschlich, Lukas Burgholzer, Robert Wille

TL;DR

This work addresses the challenge of selecting optimal quantum circuit compilation options (qubit technology, device, compiler, and settings) for a given circuit. It formulates the problem as a supervised classification task and builds a training pipeline using $3000$ circuits from MQT Bench to predict the best option with a Random Forest classifier, achieving best options for over $75 ext{ of unseen circuits}$ and top-3 for over $95 ext{ of unseen circuits}$, while reducing median compilation time by more than an order of magnitude. The framework yields interpretable knowledge via feature importance and is publicly available as MQTPredictor within the Munich Quantum Toolkit, enabling end-users to deploy quantum circuits on real hardware with guidance and rapid feedback. The results demonstrate a practical, scalable approach to automate and explain the selection of compilation options, potentially accelerating adoption of quantum computing across diverse domains.

Abstract

Any potential application of quantum computing, once encoded as a quantum circuit, needs to be compiled in order to be executed on a quantum computer. Deciding which qubit technology, which device, which compiler, and which corresponding settings are best for the considered problem -- according to a measure of goodness -- requires expert knowledge and is overwhelming for end-users from different domains trying to use quantum computing to their advantage. In this work, we treat the problem as a statistical classification task and explore the utilization of supervised machine learning techniques to optimize the compilation of quantum circuits. Based on that, we propose a framework that, given a quantum circuit, predicts the best combination of these options and, therefore, automatically makes these decisions for end-users. Experimental evaluations show that, considering a prototypical setting with 3000 quantum circuits, the proposed framework yields promising results: for more than three quarters of all unseen test circuits, the best combination of compilation options is determined. Moreover, for more than 95% of the circuits, a combination of compilation options within the top-three is determined -- while the median compilation time is reduced by more than one order of magnitude. Furthermore, the resulting methodology not only provides end-users with a prediction of the best compilation options, but also provides means to extract explicit knowledge from the machine learning technique. This knowledge helps in two ways: it lays the foundation for further applications of machine learning in this domain and, also, allows one to quickly verify whether a machine learning algorithm is reasonably trained. The corresponding framework and the pre-trained classifier are publicly available on GitHub (https://github.com/cda-tum/MQTPredictor) as part of the Munich Quantum Toolkit (MQT).

Predicting Good Quantum Circuit Compilation Options

TL;DR

This work addresses the challenge of selecting optimal quantum circuit compilation options (qubit technology, device, compiler, and settings) for a given circuit. It formulates the problem as a supervised classification task and builds a training pipeline using circuits from MQT Bench to predict the best option with a Random Forest classifier, achieving best options for over and top-3 for over , while reducing median compilation time by more than an order of magnitude. The framework yields interpretable knowledge via feature importance and is publicly available as MQTPredictor within the Munich Quantum Toolkit, enabling end-users to deploy quantum circuits on real hardware with guidance and rapid feedback. The results demonstrate a practical, scalable approach to automate and explain the selection of compilation options, potentially accelerating adoption of quantum computing across diverse domains.

Abstract

Any potential application of quantum computing, once encoded as a quantum circuit, needs to be compiled in order to be executed on a quantum computer. Deciding which qubit technology, which device, which compiler, and which corresponding settings are best for the considered problem -- according to a measure of goodness -- requires expert knowledge and is overwhelming for end-users from different domains trying to use quantum computing to their advantage. In this work, we treat the problem as a statistical classification task and explore the utilization of supervised machine learning techniques to optimize the compilation of quantum circuits. Based on that, we propose a framework that, given a quantum circuit, predicts the best combination of these options and, therefore, automatically makes these decisions for end-users. Experimental evaluations show that, considering a prototypical setting with 3000 quantum circuits, the proposed framework yields promising results: for more than three quarters of all unseen test circuits, the best combination of compilation options is determined. Moreover, for more than 95% of the circuits, a combination of compilation options within the top-three is determined -- while the median compilation time is reduced by more than one order of magnitude. Furthermore, the resulting methodology not only provides end-users with a prediction of the best compilation options, but also provides means to extract explicit knowledge from the machine learning technique. This knowledge helps in two ways: it lays the foundation for further applications of machine learning in this domain and, also, allows one to quickly verify whether a machine learning algorithm is reasonably trained. The corresponding framework and the pre-trained classifier are publicly available on GitHub (https://github.com/cda-tum/MQTPredictor) as part of the Munich Quantum Toolkit (MQT).
Paper Structure (18 sections, 6 figures, 1 table)

This paper contains 18 sections, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Considered Problem: Determining Good Compilation Options
  3. Motivation
  4. Related Work
  5. Proposed Solution: Predicting Good Compilation Options
  6. Compilation Options and Compilation Pipeline
  7. Evaluation Metric
  8. Generation of Training Data
  9. Solving the Classification Task
  10. Resulting Prediction Framework and its Performance
  11. Setting Up the Prediction Framework
  12. Resulting Performance
  13. Importance of Compilation Options
  14. Discussion
  15. Knowledge Extraction
  16. ...and 3 more sections

Figures (6)

  • Figure 1: Search space of compilation options for the setup described in \ref{['ex:compilation_paths']}.
  • Figure 2: Generation of a training circuit from an example circuit.
  • Figure 3: Comparison of supervised machine learning classifiers.
  • Figure 4: Detailed evaluation of all combinations of compilation options for all unseen test circuits. All unseen test circuits are sorted by their qubit number which is indicated on the x-axis and, for most qubit numbers, multiple test circuits are evaluated. For each test circuit, all executed and evaluated combinations of compilation options are indicated by green dots () while the predicted combination of compilation options is indicated by a purple dot (). Due to the increased availability of quantum circuits with small numbers of qubits, the test circuits are not evenly spaced out with regard to their number of qubits. Since each device comes with a qubit limit (e.g., $8$, $11$, $27$, $80$, and $127$ qubits), the number of executed combinations of compilation options (and corresponding green dots) decreases with a growing number of qubits. For the majority of the test cases, the predicted combination of compilation options returns the best results---illustrating the performance of the proposed framework.
  • Figure 5: Simplified decision tree classifier.
  • ...and 1 more figures

Theorems & Definitions (6)

  • Example 1
  • Example 2
  • Example 3
  • Example 4
  • Example 5
  • Example 6