DeepAI AI Chat
Log In Sign Up

Testing MCMC code

12/16/2014
by   Roger B. Grosse, et al.
UNIVERSITY OF TORONTO
Harvard University
0

Markov Chain Monte Carlo (MCMC) algorithms are a workhorse of probabilistic modeling and inference, but are difficult to debug, and are prone to silent failure if implemented naively. We outline several strategies for testing the correctness of MCMC algorithms. Specifically, we advocate writing code in a modular way, where conditional probability calculations are kept separate from the logic of the sampler. We discuss strategies for both unit testing and integration testing. As a running example, we show how a Python implementation of Gibbs sampling for a mixture of Gaussians model can be tested.

READ FULL TEXT

page 1

page 2

page 3

page 4

01/17/2020

Unit Testing for MCMC and other Monte Carlo Methods

We propose approaches for testing implementations of Markov Chain Monte ...
11/24/2016

Quantum Enhanced Inference in Markov Logic Networks

Markov logic networks (MLNs) reconcile two opposing schools in machine l...
07/20/2020

Automating Involutive MCMC using Probabilistic and Differentiable Programming

Involutive MCMC is a unifying mathematical construction for MCMC kernels...
01/16/2013

Minimum Message Length Clustering Using Gibbs Sampling

The K-Mean and EM algorithms are popular in clustering and mixture model...
11/18/2019

emcee v3: A Python ensemble sampling toolkit for affine-invariant MCMC

emcee is a Python library implementing a class of affine-invariant ensem...
10/23/2019

Automatically Batching Control-Intensive Programs for Modern Accelerators

We present a general approach to batching arbitrary computations for acc...