manic-xai 1.0.68

Genetic Algorithm for Generating Metacounterfactual Explanations

\n# Manic: A Genetic Algorithm-based Metaheuristic Approach for Nature-Inspired Aggregation of Counterfactuals

Table of Contents

• Introduction
• Installation
• Usage
• Docker Instructions
• Citation
• Author
• Contact
• License

Introduction

Manic is a Python package that provides a genetic algorithm-based metaheuristic approach for aggregating counterfactual explanations. It implements a nature-inspired optimization technique to generate counterfactuals that explain the disagreement between different explainers. The goal is to find counterfactual instances that are both diverse and informative to enhance the interpretability of machine learning models.

Installation

To install Manic, use the following pip3 command:

pip3 install manic-xai

Usage

You can use Manic in your Python code as follows:

from manic import Manic

# Define your data_instance, base_counterfactuals, categorical_features, immutable_features, feature_ranges, data, and predict_fn

manic_instance = Manic(data_instance, base_counterfactuals, categorical_features, immutable_features, feature_ranges, data, predict_fn)

# Generate counterfactuals
counterfactuals = manic_instance.generate_counterfactuals()

Docker Instructions

To run Manic using Docker, follow these steps:

1. Build the Docker image:

docker build -t manic .

2. Run the Docker container

docker run -v /path/to/your/data:/data -it manic python3 your_script.py

Citation

If you use the Manic package in your research or work and find it helpful, we kindly request that you cite it using the following BibTeX entry:

@software{manic,
  author       = {Craig Pirie},
  title        = {Manic: A Genetic Algorithm-based Metaheuristic Approach for Nature-Inspired Aggregation of Counterfactuals},
  year         = {2023},
  publisher    = {GitHub},
  journal      = {GitHub repository},
  howpublished = {\url{https://github.com/your-username/manic}},
}

We appreciate your support and acknowledgment of our work.

Contact

For any inquiries or collaborations, please contact Craig Pirie at [e-mail].

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contributions

Contributions to the Manic package are always welcome. If you find any issues or have ideas for improvements, please feel free to open an issue or submit a pull request on the GitHub repository. Together, we can make Manic better for everyone.

Changelog

For updates and a history of changes to the Manic package, please refer to the Changelog.