Deep probabilistic programming combines deep neural networks (for automatic
hierarchical representation learning) with probabilistic models (for principled
handling of uncertainty). Unfortunately, it is difficult to write deep
probabilistic models, because existing programming frameworks lack concise,
high-level, and clean ways to express them. To ease this task, we extend Stan,
a popular high-level probabilistic programming language, to use deep neural
networks written in PyTorch. Training deep probabilistic models works best with
variational inference, so we also extend Stan for that. We implement these
extensions by translating Stan programs to Pyro. Our translation clarifies the
relationship between different families of probabilistic programming languages.
Overall, our paper is a step towards making deep probabilistic programming
easier.
Description
[1810.00873] Extending Stan for Deep Probabilistic Programming
%0 Generic
%1 burroni2018extending
%A Burroni, Javier
%A Baudart, Guillaume
%A Mandel, Louis
%A Hirzel, Martin
%A Shinnar, Avraham
%D 2018
%K bayesian deep_learning probabilistic_programming statistics
%T Extending Stan for Deep Probabilistic Programming
%U http://arxiv.org/abs/1810.00873
%X Deep probabilistic programming combines deep neural networks (for automatic
hierarchical representation learning) with probabilistic models (for principled
handling of uncertainty). Unfortunately, it is difficult to write deep
probabilistic models, because existing programming frameworks lack concise,
high-level, and clean ways to express them. To ease this task, we extend Stan,
a popular high-level probabilistic programming language, to use deep neural
networks written in PyTorch. Training deep probabilistic models works best with
variational inference, so we also extend Stan for that. We implement these
extensions by translating Stan programs to Pyro. Our translation clarifies the
relationship between different families of probabilistic programming languages.
Overall, our paper is a step towards making deep probabilistic programming
easier.
@misc{burroni2018extending,
abstract = {Deep probabilistic programming combines deep neural networks (for automatic
hierarchical representation learning) with probabilistic models (for principled
handling of uncertainty). Unfortunately, it is difficult to write deep
probabilistic models, because existing programming frameworks lack concise,
high-level, and clean ways to express them. To ease this task, we extend Stan,
a popular high-level probabilistic programming language, to use deep neural
networks written in PyTorch. Training deep probabilistic models works best with
variational inference, so we also extend Stan for that. We implement these
extensions by translating Stan programs to Pyro. Our translation clarifies the
relationship between different families of probabilistic programming languages.
Overall, our paper is a step towards making deep probabilistic programming
easier.},
added-at = {2018-12-07T09:43:40.000+0100},
author = {Burroni, Javier and Baudart, Guillaume and Mandel, Louis and Hirzel, Martin and Shinnar, Avraham},
biburl = {https://www.bibsonomy.org/bibtex/20f3993617d3cc835ade6788cbb4c0880/jpvaldes},
description = {[1810.00873] Extending Stan for Deep Probabilistic Programming},
interhash = {aa3208e5de89a665df4e8a6e3512d53d},
intrahash = {0f3993617d3cc835ade6788cbb4c0880},
keywords = {bayesian deep_learning probabilistic_programming statistics},
note = {cite arxiv:1810.00873},
timestamp = {2018-12-07T09:43:40.000+0100},
title = {Extending Stan for Deep Probabilistic Programming},
url = {http://arxiv.org/abs/1810.00873},
year = 2018
}