quantumopt

quantumopt is an advanced AI-driven quantum circuit compiler that leverages Graph Neural Networks and the Claude LLM for optimizing quantum circuits specifically for IBM quantum hardware. This innovative tool transforms your quantum circuit into an optimized version that not only improves run accuracy but also provides a comprehensive explanation that can be cited in academic works.

Key Features

• Optimal Circuit Optimization: Enhances circuits based on real research data, improving performance on IBM's quantum hardware.
• Predictive Optimization Potential: Quickly assesses the potential gains from optimization, allowing researchers to prioritize circuit processing effectively.

Benefits of Predicting Optimization Potential

Utilizing quantumopt streamlines the optimization process, providing significant insights:

1. Prioritization: Identify which circuits will benefit most from optimization efforts.
2. Budget Allocation: Optimize resource utilization by intelligently managing compute budgets.
3. Bypass Ineffective Optimizations: Save time by skipping circuits that do not require optimization, avoiding unnecessary overhead.
4. Rapid Estimation: Obtain swift estimates before full transpilation, facilitating faster decision-making.

The Graph Neural Network (GNN) in quantumopt predicts optimization potential in under a second, significantly faster than the standard 3-10 seconds required for full compilation.

Usage

Begin using quantumopt with a simple integration into your quantum circuits:

from qiskit import QuantumCircuit
from quantumopt import compile

# Define your quantum research circuit
qc = QuantumCircuit(5)
qc.h(0)
cx(0, 1)
cx(1, 2)
qc.ry(0.5, 0)
qc.rz(0.3, 1)

# Compile and optimize the circuit
result = compile(qc, hardware="ibm_brisbane")

# Display optimization results
print(result.depth_reduction)    # e.g., "31%"
print(result.gate_reduction)     # e.g., "32%"
print(result.explanation)        # Detailed report from Claude
print(result.optimized_circuit)  # Optimized circuit ready for IBM

Benchmark Performance

Quantumopt has been rigorously tested on both real and synthetic circuits:

Real Circuit Testing (41 Circuits from QASMbench)

Synthetic Circuit Testing (10,240 Circuits)

Explanation Output

When configured with the ANTHROPIC_API_KEY, quantumopt generates detailed explanations:

"Transpilation of the target circuit for IBM Brisbane hardware yielded a 31.6% reduction in circuit depth (128 → 88 layers) and a 31.6% reduction in total gate count (326 → 223 gates)."

This quality of explanation aids in understanding the benefits of the optimizations applied, making it suitable for academic citations.

Supported Algorithms and Hardware

• Optimized Algorithms: VQE, QAOA, QFT, Grover, and more.
• Compatible Hardware: Primarily designed for IBM Brisbane, with additional backends planned for future releases.

Requirements

To utilize quantumopt, the following are necessary:

• Python 3.10+
• Qiskit version 1.0.0 or higher
• PyTorch version 2.0.0 or higher
• torch-geometric version 2.4.0 or higher
• optional: anthropic package for enhanced explanation generation

For those in the research community, citation of quantumopt should reference:

Syamala, N. (2025). quantumopt: An AI-driven quantum circuit compiler using Graph Neural Networks and Large Language Models. GitHub. https://github.com/nsyamala1/quantumopt

For further assistance or issue reporting, users can reach out on the project's GitHub page.