pyRiemann-qiskit: Quantum Machine Learning for BCI¶

A powerful Qiskit wrapper for pyRiemann that brings quantum computing to Riemannian geometry-based brain-computer interfaces.

🚀 Quick Start

Get started with pyRiemann-qiskit in minutes. Install and run your first quantum classifier.

Installing pyRiemann-qiskit

📚 Examples

Explore our gallery of examples showcasing quantum classification with EEG/MEG data.

Examples Gallery

🔧 API Reference

Complete API documentation for all classes and functions.

API reference

Overview¶

pyRiemann-qiskit is a Qiskit wrapper around pyRiemann that enables quantum classification with Riemannian geometry for brain-computer interface applications.

Key Features¶

✨ Quantum Algorithms

Leverage quantum computing for pattern recognition and classification tasks.

🧠 Brain Signal Processing

Specialized tools for EEG/MEG data analysis using Riemannian geometry.

🔬 Research Sandbox

Experiment with quantum machine learning in a flexible environment.

⚡ Backend Flexibility

Run on local simulators, remote simulators, or real quantum hardware.

🎯 Quantum Classifiers

Quantum Support Vector Machines (QSVM)
Variational Quantum Classifiers (VQC)
Quantum Minimum Distance to Mean (MDM)
Nearest Convex Hull (NCH) algorithms

Use Cases¶

A typical workflow involves:

Preprocessing: Extract covariance matrices from EEG/MEG signals
Tangent Space: Project matrices to tangent space for vectorization
Quantum Encoding: Encode features into quantum states
Classification: Use quantum algorithms for pattern recognition

from pyriemann_qiskit.classification import QuanticSVM
from pyriemann.estimation import Covariances
from pyriemann.tangentspace import TangentSpace
from sklearn.pipeline import make_pipeline

# Create a quantum classification pipeline
clf = make_pipeline(
    Covariances(),
    TangentSpace(),
    QuanticSVM()
)

# Train and predict
clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)

Featured Examples¶

Quantum SVM for EEG

Classify EEG signals using quantum support vector machines.

ERP EEG decoding with Quantum Classifier.

Distance Visualization

Visualize distances between classes using MDM and NCH estimators.

Visualize distances with MDM and NCH

Classifier Comparison

Compare quantum vs classical classifiers on toy datasets.

Comparison with toys datasets.

Important Considerations¶

Quantum Limitations

Feature Dimension Limits

Local simulator: ~36 qubits
Real quantum hardware: up to 156 qubits

Time Complexity

Quantum algorithms may take longer on local machines. Use remote backends for better performance. Suitable for offline analysis only.

Getting Help

📖 Read the introduction for background
💻 Check the installation guide
🎨 Browse the example gallery
🐛 Report bugs on GitHub

Citation¶

If you use pyRiemann-qiskit in your research, please cite:

@article{andreev2023pyriemann,
  title={pyRiemann-qiskit: A Sandbox for Quantum Classification
         Experiments with Riemannian Geometry},
  author={Andreev, Anton and Cattan, Gr{\'e}goire and
          Chevallier, Sylvain and Barth{\'e}lemy, Quentin},
  journal={Research Ideas and Outcomes},
  volume={9},
  year={2023},
  publisher={Pensoft Publishers},
  doi={10.3897/rio.9.e101006}
}