Fetching the paper…

Reading the bibliography…

2015

Particle Gibbs with Ancestor Sampling for Probabilistic Programs

van de Meent, Jan-Willem, Yang, Hongseok, Mansinghka, Vikash et al.

Original
PDF
Post on X

Understand

Particle Markov chain Monte Carlo techniques rank among current state-of-the-art methods for probabilistic program inference.

A drawback of these techniques is that they rely on importance resampling, which results in degenerate particle trajectories and a low effective sample size for variables sampled early in a program.
We here develop a formalism to adapt ancestor resampling, a technique that mitigates particle degeneracy, to the probabilistic programming setting.
We present empirical results that demonstrate nontrivial performance gains.

Particle Gibbs with Ancestor Sampling for Probabilistic Programs · Around

Built on

Sequential Monte Carlo samplers
Pierre Del Moral, Arnaud Doucet, and Ajay Jasra · 2006
Earlier work this paper cites.
Church: a language for generative models
Noah Goodman, Vikash Mansinghka, Daniel M Roy, Keith Bonawitz, and Joshua B Tenenbaum · 2008
Earlier work this paper cites.
Particle Markov chain Monte Carlo methods
Christophe Andrieu, Arnaud Doucet, and Roman Holenstein · 2010
Earlier work this paper cites.
Infer .NET 2.4, Microsoft Research Cambridge, 2010
T Minka, J Winn, J Guiver, and D Knowles · 2010
Earlier work this paper cites.

Similar

Lightweight implementations of probabilistic programming languages via transformational compilation
David Wingate, Andreas Stuhlmueller, and Noah D Goodman · 2011
Cited alongside, same era.
Ancestor Sampling for Particle Gibbs
Fredrik Lindsten, Michael I Jordan, and Thomas B. Schön · 2012
Cited alongside, same era.
Path storage in the particle filter
Pierre E. Jacob, Lawrence M. Murray, and Sylvain Rubenthaler · 2013
Cited alongside, same era.
Venture: a higher-order probabilistic programming platform with programmable inference
Original
Vikash Mansinghka, Daniel Selsam, and Yura Perov · 2014
Cited alongside, same era.

Then

A Compilation Target for Probabilistic Programming Languages
Brooks Paige and Frank Wood · 2014
Later among the works it cites.
Stan: A C++ library for probability and sampling, version 2.5.0, 2014
Stan Development Team · 2014
Later among the works it cites.
A new approach to probabilistic programming inference
F Wood, JW van de Meent, and V Mansinghka · 2014
Later among the works it cites.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…