PPL Bench: Evaluation Framework For Probabilistic Programming Languages

TL;DR

PPL Bench is a comprehensive evaluation framework for probabilistic programming languages, enabling standardized assessment of accuracy and convergence speed across models and implementations.

Contribution

It introduces a publicly available benchmark with evaluation tools and encourages community contributions to improve PPL assessment and selection.

Findings

Provides a standardized way to evaluate PPLs

Includes metrics like effective sample size and $\hat{r}$

Facilitates comparison of accuracy and speed of convergence

Abstract

We introduce PPL Bench, a new benchmark for evaluating Probabilistic Programming Languages (PPLs) on a variety of statistical models. The benchmark includes data generation and evaluation code for a number of models as well as implementations in some common PPLs. All of the benchmark code and PPL implementations are available on Github. We welcome contributions of new models and PPLs and as well as improvements in existing PPL implementations. The purpose of the benchmark is two-fold. First, we want researchers as well as conference reviewers to be able to evaluate improvements in PPLs in a standardized setting. Second, we want end users to be able to pick the PPL that is most suited for their modeling application. In particular, we are interested in evaluating the accuracy and speed of convergence of the inferred posterior. Each PPL only needs to provide posterior samples given a model and observation data. The framework automatically computes and plots growth in predictive log-likelihood on held out data in addition to reporting other common metrics such as effective sample size and $\hat{r}$.