Inference Compilation and Universal Probabilistic Programming

TL;DR

This paper presents a novel approach called inference compilation that uses neural networks to efficiently perform approximate inference in universal probabilistic programming models, combining probabilistic programming with deep learning.

Contribution

It introduces a framework that transforms probabilistic programs into trained neural networks for fast inference, enabling significant speedups in complex models.

Findings

Achieved substantial speedups in inference efficiency.

Successfully applied to mixture models and Captcha solving.

Demonstrated the effectiveness of neural network proposals in probabilistic inference.

Abstract

We introduce a method for using deep neural networks to amortize the cost of inference in models from the family induced by universal probabilistic programming languages, establishing a framework that combines the strengths of probabilistic programming and deep learning methods. We call what we do "compilation of inference" because our method transforms a denotational specification of an inference problem in the form of a probabilistic program written in a universal programming language into a trained neural network denoted in a neural network specification language. When at test time this neural network is fed observational data and executed, it performs approximate inference in the original model specified by the probabilistic program. Our training objective and learning procedure are designed to allow the trained neural network to be used as a proposal distribution in a sequential importance sampling inference engine. We illustrate our method on mixture models and Captcha solving and show significant speedups in the efficiency of inference.