Automatic Sampler Discovery via Probabilistic Programming and Approximate Bayesian Computation

Perov, Yura; Wood, Frank

doi:10.1007/978-3-319-41649-6_27

Cited by 5 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…

This abstract extends on the previous work [21,22] on program induction [16] using probabilistic programming. It describes possible further steps to extend that work, such that, ultimately, automatic probabilistic program synthesis can generalise over any reasonable set of inputs and outputs, in particular in regard to text, image and video data.

…”

mentioning

confidence: 55%

“…In the previous work [21,22] it was shown how to define an adaptor [6], strongly-typed grammar as a probabilistic program, which can generate other probabilistic programs given a specification in the form of a set of observations: By performing inference over model-s, it was possible to infer simple probabilistic programs (specifically, samplers from one dimensional distributions, e.g. Bernoulli, Poisson, etc.)…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Inference Over Programs That Make Predictions

Perov

2018

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…

…”

mentioning

confidence: 55%

Section: Introductionmentioning

confidence: 99%

Inference Over Programs That Make Predictions

Perov

2018

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…One approach is to use a probabilistic programming system and inference to invert a generative model that generates normal, regular, computer program code and conditions on its output, when run on examples, conforming to the observed specification. This is the central idea in the work of Perov and Wood (2016) whose use of probabilistic programming is what distinguishes their work from the related literature (Gulwani et al, 2017;Hwang et al, 2011;Liang et al, 2010). Examples such as this, even more than the preceding visually compelling examples, illustrate the denotational convenience of a rich and expressive programming language as the generative modeling language.…”

Section: Program Inductionmentioning

confidence: 99%

An Introduction to Probabilistic Programming

van de Meent,

Paige,

Yang

et al. 2018

Preprint

Self Cite

View full text Add to dashboard Cite

2an interface between program executions and an inference controller.This document closes with a chapter on advanced topics which we believe to be, at the time of writing, interesting directions for probabilistic programming research; directions that point towards a tight integration with deep neural network research and the development of systems for next-generation artificial intelligence applications.We would like to thank the very large number of people who have read through preliminary versions of this manuscript. Comments from the reviewers have been particularly helpful, as well as general interactions with David Blei and Kevin Murphy in particular. Some people we would like to individually thank are, in no particular order,

show abstract

“…In future work, we would like to create a belief propagation-based back-end for TerpreT. The problem of inducing samplers for probability distributions has also been cast as a problem of inference in a probabilistic program (Perov and Wood, 2016). Lake et al ( 2015) induce probabilistic programs by performing inference in a probabilistic model describing how primitives are composed to form types and instances of types.…”

Section: Probabilistic Programming and Graphical Modelsmentioning

confidence: 99%

TerpreT: A Probabilistic Programming Language for Program Induction

Gaunt,

Brockschmidt,

Singh

et al. 2016

Preprint

View full text Add to dashboard Cite

We study machine learning formulations of inductive program synthesis; that is, given input-output examples, we would like to synthesize source code that maps inputs to corresponding outputs. Our aims in this work are to develop new machine learning approaches to the problem based on neural networks and graphical models, and to understand the capabilities of machine learning techniques relative to traditional alternatives, such as those based on constraint solving from the programming languages community.Our key contribution is the proposal of TerpreT, a domain-specific language for expressing program synthesis problems. TerpreT is similar to a probabilistic programming language: a model is composed of a specification of a program representation (declarations of random variables) and an interpreter that describes how programs map inputs to outputs (a model connecting unknowns to observations). The inference task is to observe a set of input-output examples and infer the underlying program. TerpreT has two main benefits. First, it enables rapid exploration of a range of domains, program representations, and interpreter models. Second, it separates the model specification from the inference algorithm, allowing proper like-to-like comparisons between different approaches to inference. From a single TerpreT specification we can automatically perform inference using four different back-ends that include machine learning and program synthesis approaches. These are based on gradient descent (thus each specification can be seen as a differentiable interpreter), linear program (LP) relaxations for graphical models, discrete satisfiability solving, and the Sketch program synthesis system.We illustrate the value of TerpreT by developing several interpreter models and performing an extensive empirical comparison between alternative inference algorithms on a variety of program models. Our key, and perhaps surprising, empirical finding is that constraint solvers dominate the gradient descent and LP-based formulations. We conclude with some suggestions on how the machine learning community can make progress on program synthesis.

show abstract

Automatic Sampler Discovery via Probabilistic Programming and Approximate Bayesian Computation

Cited by 5 publications

References 7 publications

Inference Over Programs That Make Predictions

Inference Over Programs That Make Predictions

An Introduction to Probabilistic Programming

TerpreT: A Probabilistic Programming Language for Program Induction

Contact Info

Product

Resources

About