Bayesian Quadrature

This library provides code for evaluating the integral of non-negative functions using Bayesian Quadrature, both serially and in a batched mode.

For some background on Bayesian quadrature, see:

• Introductory slides by David Duvenaud, or these slides by Roman Garnett

And for gory detail:

• Sampling for Inference in Probabilistic Models with Fast Bayesian Quadrature, Gunter et al. 2014 for details on the warped Gaussian Process model implemented here ("WSABI")
• Batch Selection for Parallelisation of Bayesian Quadrature for details on our batch selection process

Installation

Check out this repository and run pip install . in the root directory of the repository (i.e. in the directory containing setup.py). You should then be able to run the example scripts.

Documentation

Documentation is still a work in progress, but some docs are available at https://OxfordML.github.io/bayesquad