Skip to content

jelli - JAX-based EFT Likelihoods

jelli is a Python package for building and evaluating likelihood functions in the Effective Field Theory (EFT) framework.

Key Features

  • EFT Framework: Construction of likelihoods in EFTs, such as the Standard Model Effective Field Theory (SMEFT) and Weak Effective Theory (WET).
  • Flexibility: Supports arbitrary observable predictions provided in the POPxf data format, and a multitude of experimental likelihood assumptions.
  • JAX Integration: Built on JAX for high-performance numerical computing.
  • Differentiable: Fully differentiable likelihood functions due to JAX's autodiff, enabling efficient gradient and Hessian computations, gradient-based optimization and sampling, and more.
  • Fast: Utilizes JAX's Just-In-Time (JIT) compilation for optimized performance.
  • Multi-scale: Interfaced with rgevolve for fast renormalization group evolution using the evolution matrix formalism.

Installation

The package can be installed via pip:

pip install jelli

Repository

The source code is available on GitHub.

Citation

A paper describing jelli is in preparation.

Bugs and feature requests

Please report bugs and request features via the GitHub issues page.

Contributors

Authors:

  • Aleks Smolkovic (@alekssmolkovic)
  • Peter Stangl (@peterstangl)

License

jelli is licensed under the MIT License.