quantum-xai 0.1.1

Explainable Quantum Machine Learning Library

Quantum-XAI: Explainable Quantum Machine Learning Library

Quantum-XAI is a comprehensive, production-ready toolkit for explainable quantum machine learning. It provides a suite of quantum-native explainability methods, quantum neural network models, advanced visualizations, benchmarking tools, and research-grade features to interpret and analyze quantum neural network decisions.

---

Features

• Quantum Neural Network Models

  • Variational Quantum Classifier (VQC) implementation using PennyLane
  • Supports multiple quantum data encodings: Angle, Amplitude, and IQP encoding

• Explainability Methods

  • Quantum SHAP Explainer: SHAP-like sampling-based explanations
  • Quantum Gradient Explainer: Gradient-based explanations using parameter-shift rule
  • Quantum LIME Explainer: LIME-like local surrogate model explanations
  • Quantum Perturbation Explainer: Feature occlusion based explanations

• Visualization Tools

  • Feature importance bar charts
  • Side-by-side explanation method comparisons
  • Quantum circuit diagrams with explanation overlays
  • Radar charts for quantum feature importance

• Datasets & Utilities

  • Preprocessed quantum-ready datasets: Iris, Wine, Breast Cancer
  • Dataset loaders with normalization and binary classification options

• Benchmarking & Evaluation

  • Compare multiple explainers on test samples
  • Compute explanation consistency and quality metrics
  • Faithfulness, sparsity, stability, and top feature importance analysis

• Research Extensions

  • Quantum Fisher Information matrix computation
  • Quantum entanglement contribution analysis
  • Quantum feature interaction analysis beyond classical correlations

• Save/Load Functionality

  • Save trained models and explanations to JSON
  • Load models and explanations from JSON for reproducibility

• Complete Demo

  • End-to-end demonstration of training, explaining, visualizing, benchmarking, and reporting

---

Installation

Ensure you have Python 3.7+ installed. Install required dependencies:

pip install pennylane numpy matplotlib seaborn scikit-learn pandas

---

Usage

Quick Start Demo

Run the complete demonstration with the Iris dataset:

from quantum_xai import QuantumXAIDemo

demo = QuantumXAIDemo()
results = demo.run_complete_demo(dataset='iris', n_samples=80)

Custom Model Training and Explanation

from quantum_xai import QuantumNeuralNetwork, QuantumSHAPExplainer, QuantumGradientExplainer, QuantumXAIVisualizer
from sklearn.model_selection import train_test_split
from sklearn.datasets import load_iris

# Load data
data = load_iris()
X = data.data
y = data.target
feature_names = data.feature_names
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# Create and train model
model = QuantumNeuralNetwork(n_features=X.shape[1], n_qubits=4, n_layers=2)
model.train(X_train, y_train, epochs=100, lr=0.1)

# Create explainers
shap_explainer = QuantumSHAPExplainer(model, X_train)
gradient_explainer = QuantumGradientExplainer(model)

# Generate explanation for a sample
explanation = shap_explainer.explain(X_test, 0)

# Visualize explanation
visualizer = QuantumXAIVisualizer()
fig = visualizer.plot_feature_importance(explanation, feature_names)
fig.show()

---

Research Applications

• Benchmark quantum vs classical explainability methods
• Analyze quantum Fisher information and entanglement effects
• Extend to other quantum platforms (Qiskit, Cirq)
• Develop advanced quantum-specific explanation metrics
• Apply to real quantum datasets in chemistry, finance, and more

---

Project Structure

• QuantumNeuralNetwork: Variational quantum classifier model
• QuantumExplainer and subclasses: Explainability methods (SHAP, Gradient, LIME, Perturbation)
• QuantumXAIVisualizer: Visualization utilities
• QuantumDatasetLoader: Dataset loading and preprocessing
• QuantumXAIBenchmark: Benchmarking and evaluation tools
• QuantumXAIDemo: Complete demo and example workflows
• save_model_and_explanations / load_model_and_explanations: Persistence utilities
• QuantumXAIResearch: Advanced research features

---

License

This project is open-source and available for research, publication, and industry use.

---

Contact

For questions, contributions, or collaborations, please open an issue or pull request on the GitHub repository.

---

Acknowledgments

This library builds upon PennyLane and scikit-learn, leveraging quantum computing and classical ML explainability techniques.