Deep Probabilistic Programming

Resource type
Authors/contributors
Title
Deep Probabilistic Programming
Abstract
We propose Edward, a Turing-complete probabilistic programming language. Edward defines two compositional representations---random variables and inference. By treating inference as a first class citizen, on a par with modeling, we show that probabilistic programming can be as flexible and computationally efficient as traditional deep learning. For flexibility, Edward makes it easy to fit the same model using a variety of composable inference methods, ranging from point estimation to variational inference to MCMC. In addition, Edward can reuse the modeling representation as part of inference, facilitating the design of rich variational models and generative adversarial networks. For efficiency, Edward is integrated into TensorFlow, providing significant speedups over existing probabilistic systems. For example, we show on a benchmark logistic regression task that Edward is at least 35x faster than Stan and 6x faster than PyMC3. Further, Edward incurs no runtime overhead: it is as fast as handwritten TensorFlow.
Publication
arXiv:1701.03757 [cs, stat]
Date
2017-03-07
Accessed
2019-11-27T23:15:14Z
Library Catalog
Extra
ZSCC: 0000108 arXiv: 1701.03757
Notes

Comment: Appears in International Conference on Learning Representations, 2017. A companion webpage for this paper is available at http://edwardlib.org/iclr2017

Citation
Tran, D., Hoffman, M. D., Saurous, R. A., Brevdo, E., Murphy, K., & Blei, D. M. (2017). Deep Probabilistic Programming. ArXiv:1701.03757 [Cs, Stat]. Retrieved from http://arxiv.org/abs/1701.03757
MACHINE LEARNING
PROBABILITY & STATISTICS
Methodology
Attachment
Processing time: 0.02 seconds

Graph of references

(from Zotero to Gephi via Zotnet with this script)
Graph of references