LifeFinder is a Python package and tool to estimate the probability of life on exoplanets using astrophysical parameters and machine learning. It includes a built-in neural network model trained to predict a habitability probability based on stellar and planetary features.

You can install LifeFinder via pip:

pip install lifefinder

Or clone the repository and use Docker for containerized environment:

git clone https://github.com/afnx/lifefinder.git

cd lifefinder
docker compose -f .devcontainer/compose.yaml up -d

Before starting to use LifeFinder, configure necessary parameters such as default artifact locations, model settings etc:

lifefinder configure

Once configured, you could start training models:

lifefinder train

LifeFinder downloads the latest exoplanet data from the NASA Exoplanet Archive, cleans it, and trains the neural network model.

To retrain an existing model with different hyperparameters, just add the --retrain flag:

lifefinder train --retrain

For a full list of commands and options, run:

lifefinder --help

LifeFinder does NOT use a classical ground truth for habitability. Instead, it generates a soft, continuous "habitability zone index" (HZ index) for model training, based on the stellar insolation flux received by a planet.

• pl_insol: Stellar insolation flux (in Earth units, S⊕).
  • Earth = 1.0 S⊕.
• sigma: Controls how quickly the index decays as you move away from Earth's flux.
  • Default: 0.5 (roughly the half-width of the conservative habitable zone).

Formula:

import numpy as np
def hz_index(pl_insol, sigma=0.5):
    return np.exp(-0.5 * ((pl_insol - 1.0) / sigma) ** 2)

• Planets near Earth's flux (pl_insol ≈ 1.0) have HZ index ≈ 1.0.
• Planets far from Earth's flux get HZ index ≈ 0.0.
• This creates a smooth, interpretable target for machine learning.

LifeFinder uses a neural network regressor built with PyTorch.

• Inputs: Astrophysical features (e.g., stellar flux, planet radius, orbital period, etc.)
• Output: Planet habitability (continuous, 0.0–1.0)
• Architecture: Feedforward neural network with:
  • Multiple dense layers
  • Nonlinear activation functions (e.g., ReLU, sigmoid)
  • Regularization (dropout, L2, etc.)
• Training: Supervised learning using the generated HZ index as the ground truth.

The outputs and predictions provided by LifeFinder do not represent the actual habitability of any exoplanet, nor do they guarantee the presence or absence of life. The model is designed to estimate a probability of habitability based on available astrophysical data and machine learning techniques. The accuracy and reliability of the results depend strongly on the quality, quantity, and completeness of the input data, as well as the modeling approaches used.

Do not rely solely on LifeFinder's output for scientific conclusions, mission decisions, or any application where actual confirmation of exoplanet habitability is required. The results should always be interpreted as probabilistic estimates and be used in combination with other scientific methods and expert analysis.

We welcome your contributions!

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

• Kopparapu et al., 2013, ApJ, "Habitable Zones Around Main-Sequence Stars: New Estimates"
• NASA Exoplanet Archive

This project is licensed under the Business Source License 1.1 - see the LICENSE file for details.

• Ali Fuat Numanoglu - Initial work - @afnx

• 🐛 Report Issues
• 📧 Contact Me
• ⭐ If this project helped you, please give it a star!

_{Made with ❤️ by Ali Fuat Numanoglu}