This repository contains a Python implementation of the population Monte Carlo algorithm proposed in the following paper:
Here is a description of each file:
- main.py - Source code for running the algorithm (along with other implemented samplers) on a toy example where the target distribution is a mixture of Gaussians. The code runs independent Monte Carlo simulations of the sampler, where multiprocessing package is used for parallelization. One can use this base code to replicate the results reported in the above paper.
- samplers.py - Library containing the implementation of the proposed sampler, as well as the samplers that it is compared to in the paper, i.e., controlled mixture population Monte Carlo (CMPMC), mixture population Monte Carlo (MPMC), adaptive population importance sampling (APIS), and parallel interacting Markov adaptive importance sampling (PIMAIS).
- other_funcs.py - Code for plotting a two-dimensional contour, evaluating a Gaussian mixture distribution, and for projecting onto the probability simplex. The code for projecting onto the probability simplex can be found in this repo.