Search

Probabilistic Programming Languages (PPLs) allow users to encode statistical inference problems and automatically apply an inference algorithm to solve them. Popular inference algorithms for PPLs, such as sequential Monte Carlo (SMC) and Markov chain Monte Carlo (MCMC), are built around checkpoints -- relevant events for the inference algorithm during the execution of a probabilistic program. Deciding the location of checkpoints is, in current PPLs, not done optimally. To solve this problem, we present a static analysis technique that automatically determines checkpoints in programs, relieving PPL users of this task. The analysis identifies a set of checkpoints that execute in the same order in every program run -- they are aligned. We formalize alignment, prove the correctness of the analysis, and implement the analysis as part of the higher-order functional PPL Miking CorePPL. By utilizing the alignment analysis, we design two novel inference algorithm variants: aligned SMC and aligned lightweight MCMC. We show, through real-world experiments, that they significantly improve inference execution time and accuracy compared to standard PPL versions of SMC and MCMC.

Probabilistic programming languages (PPLs) allow users to encode arbitrary inference problems, and PPL implementations provide general-purpose automatic inference for these problems. However, constructing inference implementations that are efficient enough is challenging for many real-world problems. Often, this is due to PPLs not fully exploiting available parallelization and optimization opportunities. For example, handling probabilistic checkpoints in PPLs through continuation-passing style transformations or non-preemptive multitasking -- as is done in many popular PPLs -- often disallows compilation to low-level languages required for high-performance platforms such as GPUs. To solve the checkpoint problem, we introduce the concept of PPL control-flow graphs (PCFGs) -- a simple and efficient approach to checkpoints in low-level languages. We use this approach to implement RootPPL: a low-level PPL built on CUDA and C++ with OpenMP, providing highly efficient and massively parallel SMC inference. We also introduce a general method of compiling universal high-level PPLs to PCFGs and illustrate its application when compiling Miking CorePPL -- a high-level universal PPL -- to RootPPL. The approach is the first to compile a universal PPL to GPUs with SMC inference. We evaluate RootPPL and the CorePPL compiler through a set of real-world experiments in the domains of phylogenetics and epidemiology, demonstrating up to 6x speedups over state-of-the-art PPLs implementing SMC inference.

We consider probabilistic programming for birth-death models of evolution and introduce a new widely-applicable inference method that combines an extension of the alive particle filter (APF) with automatic Rao-Blackwellization via delayed sampling. Birth-death models of evolution are an important family of phylogenetic models of the diversification processes that lead to evolutionary trees. Probabilistic programming languages (PPLs) give phylogeneticists a new and exciting tool: their models can be implemented as probabilistic programs with just a basic knowledge of programming. The general inference methods in PPLs reduce the need for external experts, allow quick prototyping and testing, and accelerate the development and deployment of new models. We show how these birth-death models can be implemented as simple programs in existing PPLs, and demonstrate the usefulness of the proposed inference method for such models. For the popular BiSSE model the method yields an increase of the effective sample size and the conditional acceptance rate by a factor of 30 in comparison with a standard bootstrap particle filter. Although concentrating on phylogenetics, the extended APF is a general inference method that shows its strength in situations where particles are often assigned zero weight. In the case when the weights are always positive, the extra cost of using the APF rather than the bootstrap particle filter is negligible, making our method a suitable drop-in replacement for the bootstrap particle filter in probabilistic programming inference.

Probabilistic programming is a programming paradigm for expressing flexible probabilistic models. Implementations of probabilistic programming languages employ a variety of inference algorithms, where sequential Monte Carlo methods are commonly used. A problem with current state-of-the-art implementations using sequential Monte Carlo inference is the alignment of program synchronization points. We propose a new static analysis approach based on the 0-CFA algorithm for automatically aligning higher-order probabilistic programs. We evaluate the automatic alignment on a phylogenetic model, showing a significant decrease in runtime and increase in accuracy., Comment: 19 pages, 10 figures

Variation in the evolutionary process across the sites of nucleotide sequence alignments is well established, and is an increasingly pervasive feature of datasets composed of gene regions sampled from multiple loci and/or different genomes. Inference of phylogeny from these data demands that we adequately model the underlying process heterogeneity; failure to do so can lead to biased estimates of phylogeny and other parameters. Traditionally, process heterogeneity has been accommodated by first assigning sites to data subsets based on relevant prior information (reflecting codon positions in protein-coding DNA, stem and loop regions of ribosomal DNA, etc.), and then estimating the phylogeny and other model parameters under the resulting mixed model. Here, we consider an alternative approach for accommodating process heterogeneity that is similar in spirit to this conventional mixed-model approach. However, rather than treating the partitioning scheme as a fixed assumption of the analysis, we treat the process partition as a random variable using a Dirichlet process prior model, where the phylogeny is estimated by integrating over all possible process partitions for the specified data subsets. We apply this method to simulated and empirical datasets, and compare our results to those estimated previously using conventional mixed-model selection criteria based on Bayes factors. We find that estimation under the Dirichlet process prior model discovers novel process partitions that may more effectively balance error variance and estimation bias, while rendering phylogenetic inference more robust to process heterogeneity by virtue of integrating estimates over all possible partition schemes. (Keywords: Bayesian phylogenetic inference; Dirichlet process prior; Markov chain Monte Carlo; maximum likelihood; partitioned analyses; process heterogeneity.)

Recent years have seen a rapid expansion of the model space explored in statistical phylogenetics, emphasizing the need for new approaches to statistical model representation and software development. Clear communication and representation of the chosen model is crucial for: (1) reproducibility of an analysis, (2) model development and (3) software design. Moreover, a unified, clear and understandable framework for model representation lowers the barrier for beginners and non-specialists to grasp complex phylogenetic models, including their assumptions and parameter/variable dependencies. Graphical modeling is a unifying framework that has gained in popularity in the statistical literature in recent years. The core idea is to break complex models into conditionally independent distributions. The strength lies in the comprehensibility, flexibility, and adaptability of this formalism, and the large body of computational work based on it. Graphical models are well-suited to teach statistical models, to facilitate communication among phylogeneticists and in the development of generic software for simulation and statistical inference. Here, we provide an introduction to graphical models for phylogeneticists and extend the standard graphical model representation to the realm of phylogenetics. We introduce a new graphical model component, tree plates, to capture the changing structure of the subgraph corresponding to a phylogenetic tree. We describe a range of phylogenetic models using the graphical model framework and introduce modules to simplify the representation of standard components in large and complex models. Phylogenetic model graphs can be readily used in simulation, maximum likelihood inference, and Bayesian inference using, for example, Metropolis-Hastings or Gibbs sampling of the posterior distribution.