Omega.jl

Omega.jl is a programming language for causal and probabilistic reasoning. It was developed by Zenna Tavares with help from Javier Burroni, Edgar Minasyan, Xin Zhang, Rajesh Ranganath and Armando Solar Lezama.

Quick Start

Omega is built in Julia 1.0. You can easily install it from a Julia repl with:

(v1.0) pkg> add Omega

Check Omega is working and gives reasonable results with:

julia> using Omega

julia> rand(normal(0.0, 1.0))
0.7625637212030862

With that, see the Tutorial for a run through of the main features of Omega.

Contribute

We want your contributions!

Probabilistic models

Please add probabilistic models and model families to https://github.com/zenna/OmegaModels.jl

Inference procedures

Citation

If you use Omega, please cite Omega papers:

The Random Conditional Distribution for Uncertain Distributional Properties

@article{tavares2019rcd,
  title={The Random Conditional Distribution for Uncertain Distributional Properties},
  author={Tavares, Zenna and Burroni, Javier and Minaysan, Edgar and Ranganath, Rajesh and Lezama, Armando Solar},
  journal={arXiv},
  year={2019}
}

Soft Constraints for Inference with Declarative Knowedlge

@article{tavares2019soft,
  title={Soft Constraints for Inference with Declarative Knowledge},
  author={Tavares, Zenna and Burroni, Javier and Minaysan, Edgar and Lezama, Armando Solar and Ranganath, Rajesh},
  journal={arXiv preprint arXiv:1901.05437},
  year={2019}
}

If you use the causal inference features (replace), please cite:

A Language for Counterfactual Generative Models

@article{tavares2019counterfactual,
  title={Soft Constraints for Inference with Declarative Knowledge},
  author={Tavares, Zenna and Zhang, Xin and Koppel, James and Lezama, Armando Solar},
  year={2019}
}

Acknowledgements

Omega leans heavily on the hard work of many packages and the Julia community as a whole, but in particular Distributions.jl, Flux.jl, and Cassette.jl.