Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DeQompile: quantum circuit decompilation using genetic programming for explainable quantum architecture search

Shubing Xie, Aritra Sarkar, Sebastian Feld

TL;DR

DeQompile addresses the explainability gap in quantum architecture search by reverse-engineering high-level Qiskit algorithms from low-level QASM using a genetic-programming-based decompiler. The framework leverages abstract syntax trees and symbolic regression to evolve human-readable circuit descriptions, validated on standard quantum algorithms and simple ansatz patterns. Key contributions include a complete GP-based decompilation pipeline, novel fitness metrics combining syntax, structure, and parsimony, and open-source availability for community use. The work demonstrates that explainable QAS is feasible for a range of circuits and highlights the tradeoffs between readability and hardware-efficient decompositions, with implications for integrating explainable methods into online learning frameworks and broader quantum algorithm design.

Abstract

Demonstrating quantum advantage using conventional quantum algorithms remains challenging on current noisy gate-based quantum computers. Automated quantum circuit synthesis via quantum machine learning has emerged as a promising solution, employing trainable parametric quantum circuits to alleviate this. The circuit ansatz in these solutions is often designed through reinforcement learning-based quantum architecture search when the domain knowledge of the problem and hardware are not effective. However, the interpretability of these synthesized circuits remains a significant bottleneck, limiting their scalability and applicability across diverse problem domains. This work addresses the challenge of explainability in quantum architecture search (QAS) by introducing a novel genetic programming-based decompiler framework for reverse-engineering high-level quantum algorithms from low-level circuit representations. The proposed approach, implemented in the open-source tool DeQompile, employs program synthesis techniques, including symbolic regression and abstract syntax tree manipulation, to distill interpretable Qiskit algorithms from quantum assembly language. Validation of benchmark algorithms demonstrates the efficacy of our tool. By integrating the decompiler with online learning frameworks, this research potentiates explainable QAS by fostering the development of generalizable and provable quantum algorithms.

DeQompile: quantum circuit decompilation using genetic programming for explainable quantum architecture search

TL;DR

DeQompile addresses the explainability gap in quantum architecture search by reverse-engineering high-level Qiskit algorithms from low-level QASM using a genetic-programming-based decompiler. The framework leverages abstract syntax trees and symbolic regression to evolve human-readable circuit descriptions, validated on standard quantum algorithms and simple ansatz patterns. Key contributions include a complete GP-based decompilation pipeline, novel fitness metrics combining syntax, structure, and parsimony, and open-source availability for community use. The work demonstrates that explainable QAS is feasible for a range of circuits and highlights the tradeoffs between readability and hardware-efficient decompositions, with implications for integrating explainable methods into online learning frameworks and broader quantum algorithm design.

Abstract

Demonstrating quantum advantage using conventional quantum algorithms remains challenging on current noisy gate-based quantum computers. Automated quantum circuit synthesis via quantum machine learning has emerged as a promising solution, employing trainable parametric quantum circuits to alleviate this. The circuit ansatz in these solutions is often designed through reinforcement learning-based quantum architecture search when the domain knowledge of the problem and hardware are not effective. However, the interpretability of these synthesized circuits remains a significant bottleneck, limiting their scalability and applicability across diverse problem domains. This work addresses the challenge of explainability in quantum architecture search (QAS) by introducing a novel genetic programming-based decompiler framework for reverse-engineering high-level quantum algorithms from low-level circuit representations. The proposed approach, implemented in the open-source tool DeQompile, employs program synthesis techniques, including symbolic regression and abstract syntax tree manipulation, to distill interpretable Qiskit algorithms from quantum assembly language. Validation of benchmark algorithms demonstrates the efficacy of our tool. By integrating the decompiler with online learning frameworks, this research potentiates explainable QAS by fostering the development of generalizable and provable quantum algorithms.

Paper Structure

This paper contains 42 sections, 8 equations, 10 figures, 2 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Background
  3. Decompilation
  4. Genetic programming
  5. Abstract syntax tree
  6. Quantum circuit decompilation
  7. Problem formulation
  8. AST representation
  9. Initialization of the population
  10. Qubit index expressions
  11. Loop structures
  12. Angle expressions for rotation gates
  13. Quantum gate calls
  14. Evolution operations
  15. Fitness evaluation
  16. ...and 27 more sections

Figures (10)

  • Figure 1: Software architecture of DeQompile
  • Figure 2: AST structure for a simple Qiskit function
  • Figure 3: An example of the effect of AST crossover on Qiskit codes
  • Figure 4: DeQompile performance for 1-qubit ansatz. The mean best score and maximum score across generations demonstrate steady improvement. The parameters used for generating the plot are: mutation_rate=0.3, pop_size=40, generations=100, rep=3, total_qubit=20, max_length=10, perform_crossover=True, crossover_rate=0.3, new_gen_rate=0.2, max_loop_depth=2, mutation_rate_2=0.5
  • Figure 5: Best decompiled codes for 1-qubit ansatz
  • ...and 5 more figures