pocoMC is a Python implementation of the Preconditioned Monte Carlo method for accelerated Bayesian inference

Getting started

Brief introduction

pocoMC utilises a Normalising Flow in order to precondition the target distribution by removing any correlations between its parameters. The code then generates posterior samples, that can be used for parameter estimation, using a powerful adaptive Sequential Monte Carlo algorithm manifesting a sampling effiency that can be orders of magnitude higher than without precondition. Furthermore, pocoMC also provides an unbiased estimate of the model evidence that can be used for the task of Bayesian model comparison.

Documentation

Read the docs at pocomc.readthedocs.io for more information, examples and tutorials.

Installation

To install pocomc using pip run:

pip install pocomc

or, to install from source:

git clone https://github.com/minaskar/pocomc.git
cd pocomc
python setup.py install

Basic example

For instance, if you wanted to draw samples from a 10-dimensional Rosenbrock distribution with a uniform prior, you would do something like:

import pocomc as pc
import numpy as np

n_dim = 10  # Number of dimensions

def log_prior(x):
    if np.any((x < -10.0) | (x > 10.0)):  # If any dimension is out of bounds, the log prior is -infinity
        return -np.inf 
    else:
        return 0.0

def log_likelihood(x):
    return -np.sum(10.0*(x[:,::2]**2.0 - x[:,1::2])**2.0 \
            + (x[:,::2] - 1.0)**2.0, axis=1)


n_particles = 1000
prior_samples = np.random.uniform(size=(n_particles, n_dim), low=-10.0, high=10.0)

sampler = pc.Sampler(
    n_particles,
    n_dim,
    log_likelihood,
    log_prior,
    vectorize_likelihood=True,
    bounds=(-10.0, 10.0)
)
sampler.run(prior_samples)

results = sampler.results # Dictionary with results

Attribution & Citation

Please cite the following papers if you found this code useful in your research:

@article{karamanis2022accelerating,
    title={Accelerating astronomical and cosmological inference with preconditioned Monte Carlo},
    author={Karamanis, Minas and Beutler, Florian and Peacock, John A and Nabergoj, David and Seljak, Uro{\v{s}}},
    journal={Monthly Notices of the Royal Astronomical Society},
    volume={516},
    number={2},
    pages={1644--1653},
    year={2022},
    publisher={Oxford University Press}
}

@article{karamanis2022pocomc,
    title={pocoMC: A Python package for accelerated Bayesian inference in astronomy and cosmology},
    author={Karamanis, Minas and Nabergoj, David and Beutler, Florian and Peacock, John A and Seljak, Uros},
    journal={arXiv preprint arXiv:2207.05660},
    year={2022}
}

Licence

pocoMC is free software made available under the GPL-3.0 License. For details see the LICENSE file.

GraCosPA / pocomc