Timezone: »
We'll be demonstrating Stochastic Matlab, a new probabilistic programming language. Stochastic Matlab allows users to write generative models in Matlab (freely mixing deterministic and random primitives, MEX files, and any other Matlab construct) and then condition the model on data. Stochastic Matlab then performs inference automatically using a variety of inference methods, including standard MCMC, parallel tempering, and hamiltonian Monte Carlo. Lightweight GPU integration is also provided. Probabilistic programming languages allow users to rapidly iterate models, testing them under a variety of inference methods. The goal of Stochastic Matlab is to scale to large datasets by taking advantage of GPUs and clusters. (Note: this project is an open-source project, and is not affiliated with Mathworks.)
Author Information
David Wingate (Brigham Young University)
More from the Same Authors
-
2011 Poster: Nonstandard Interpretations of Probabilistic Programs for Efficient Inference »
David Wingate · Noah Goodman · Andreas Stuhlmueller · Jeffrey Siskind -
2010 Poster: Nonparametric Bayesian Policy Priors for Reinforcement Learning »
Finale P Doshi-Velez · David Wingate · Nicholas Roy · Josh Tenenbaum -
2007 Oral: Exponential Family Predictive Representations of State »
David Wingate · Satinder Singh -
2007 Poster: Exponential Family Predictive Representations of State »
David Wingate · Satinder Singh -
2006 Workshop: Grounding Perception, Knowledge and Cognition in Sensori-Motor Experience »
Michael James · David Wingate · Brian Tanner