Formerly: Harvard Intelligent Probabilistic Systems Group -- Now at Princeton

Efficiently computes derivatives of NumPy code.

Dependent multinomials made easy: stick-breaking with the Pólya-gamma augmentation

Spearmint Bayesian optimization codebase

Implementation in C and Theano of the method Probabilistic Backpropagation for scalable Bayesian inference in deep neural networks.

Convolutional nets which can take molecular graphs of arbitrary size as input.

Library of common tools for machine learning research.

Kayak is a library for automatic differentiation with applications to deep neural networks.

A project to enable optimization of molecules by transforming them to and from a continuous representation.

Implementation of an algorithm for Markov chain Monte Carlo with data subsampling

MCMC for the Dark Energy Spectroscopic Instrument