https://github.com/aglebov/gradient-free-mcmc-postprocessing

Gradient-free optimal postprocessing of MCMC output
https://github.com/aglebov/gradient-free-mcmc-postprocessing

Last synced: about 16 hours ago
JSON representation

Gradient-free optimal postprocessing of MCMC output

• Host: GitHub
• URL: https://github.com/aglebov/gradient-free-mcmc-postprocessing
• Owner: aglebov
• License: mit
• Created: 2024-06-02T16:21:11.000Z (5 months ago)
• Default Branch: main
• Last Pushed: 2024-09-16T22:38:03.000Z (about 2 months ago)
• Last Synced: 2024-09-18T03:34:34.271Z (about 2 months ago)
• Language: Jupyter Notebook
• Size: 85.4 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Gradient-Free Optimal Postprocessing of MCMC Output

The project aims to extend the work in 

> 1. `Riabiz, M., Chen, W. Y., Cockayne, J., Swietach, P., Niederer, S. A., Mackey, L., Oates, C. J. (2022). Optimal thinning of MCMC output. Journal of the Royal Statistical Society Series B: Statistical Methodology, 84(4), 1059-1081`.

by implementing the idea presented in

> 2. `Fisher, M. A., Oates, C. (2022). Gradient-free kernel Stein discrepancy. arXiv preprint arXiv:2207.02636.`

We replicate the results from [1] for the Lotka-Volterra model in ``code/lotka_volterra/Stein_thinning.ipynb``.

``code/notebooks/gaussian_mixture/Gaussian_mixture.ipynb`` demonstrates using gradient-free kernel Stein density as proposed in [2] for a bivariate Gaussian mixture.

The ``code/notebooks/examples`` directory also contains several examples of using the relevant Python packages.

To run the code, navigate to the `code` directory, create and activate a virtual environment and run the following command:

```

pip install -e .

```