Pyro: Deep Universal Probabilistic Programming

Abstract

Pyro is a probabilistic programming language built on Python as a platform for developing advanced probabilistic models in AI research. To scale to large datasets and high-dimensional models, Pyro uses stochastic variational inference algorithms and probability distributions built on top of PyTorch, a modern GPU-accelerated deep learning framework. To accommodate complex or model-specific algorithmic behavior, Pyro leverages Poutine, a library of composable building blocks for modifying the behavior of probabilistic programs.

Submitted to JMLR MLOSS track

References in corpus (4)

• Auto-Encoding Variational Bayes
• Church: a language for generative models
• TensorFlow Distributions
• Learning Disentangled Representations with Semi-Supervised Deep Generative Models

Cited by in corpus (76)

• The CosmoVerse White Paper: Addressing observational tensions in cosmology with systematics and fundamental physics
• Gravitational-wave parameter estimation with autoregressive neural network flows
• Mining gold from implicit models to improve likelihood-free inference
• A survey of multimodal deep generative models
• Learning Disentangled Representations in the Imaging Domain
• Scaling Exact Inference for Discrete Probabilistic Programs
• Bayesian Layers: A Module for Neural Network Uncertainty
• Uncertainty Quantification in Multivariable Regression for Material Property Prediction with Bayesian Neural Networks
• Physics-guided Deep Markov Models for Learning Nonlinear Dynamical Systems with Uncertainty
• Simple, Distributed, and Accelerated Probabilistic Programming
• Open vs Closed-ended questions in attitudinal surveys -- comparing, combining, and interpreting using natural language processing
• A Whole Brain Probabilistic Generative Model: Toward Realizing Cognitive Architectures for Developmental Robots
• Variational Bayes survival analysis for unemployment modelling
• A Tutorial on Deep Latent Variable Models of Natural Language
• Advancing Thermodynamic Group-Contribution Methods by Machine Learning: UNIFAC 2.0
• Semi-supervised source localization with deep generative modeling
• Nonlinear System Identification: Learning while respecting physical models using a sequential Monte Carlo method
• Black-Box Optimization with Local Generative Surrogates
• Combining data assimilation and machine learning to estimate parameters of a convective-scale model
• Incorporating structured assumptions with probabilistic graphical models in fMRI data analysis
• Bayesian Calibration of MEMS Accelerometers
• ADEV: Sound Automatic Differentiation of Expected Values of Probabilistic Programs
• Sine-skewed toroidal distributions and their application in protein bioinformatics
• Identification and Correction of Temporal and Spatial Distortions in Scanning Transmission Electron Microscopy
• Disentangled Representation for Causal Mediation Analysis
• A tutorial on the Bayesian statistical approach to inverse problems
• Compiling Universal Probabilistic Programming Languages with Efficient Parallel Sequential Monte Carlo Inference
• CUQIpy: I. Computational uncertainty quantification for inverse problems in Python
• On Probabilistic Termination of Functional Programs with Continuous Distributions
• Densities of Almost Surely Terminating Probabilistic Programs are Differentiable Almost Everywhere
• Exploration of lattice Hamiltonians for functional and structural discovery via Gaussian Process-based Exploration-Exploitation
• Functional Tensors for Probabilistic Programming
• A Bayesian neural network approach to Multi-fidelity surrogate modelling
• Bayesian model of electrical heating disaggregation
• Modular Probabilistic Models via Algebraic Effects
• Preconditioned FEM-based Neural Networks for Solving Incompressible Fluid Flows and Related Inverse Problems
• Disentangling transients and their host galaxies with Scarlet2: A framework to forward model multi-epoch imaging
• Bayesian Neural Networks at Scale: A Performance Analysis and Pruning Study
• Asynchronous Effects
• Map-level baryonification: unified treatment of weak lensing two-point and higher-order statistics
• Learning to solve Bayesian inverse problems: An amortized variational inference approach using Gaussian and Flow guides
• Simplifying Dependent Reductions in the Polyhedral Model
• Inferring Signaling Pathways with Probabilistic Programming
• MetFlow: A New Efficient Method for Bridging the Gap between Markov Chain Monte Carlo and Variational Inference
• Independent finite approximations for Bayesian nonparametric inference
• Statically Bounded-Memory Delayed Sampling for Probabilistic Streams
• A Constrained NILC method for CMB B mode observations
• Bayesian sparsification for deep neural networks with Bayesian model reduction
• Bayesian propensity score matching in automotive embedded software engineering
• Direct Evolutionary Optimization of Variational Autoencoders With Binary Latents
• SaberLDA: Sparsity-Aware Learning of Topic Models on GPUs
• Efficient Computation Reduction in Bayesian Neural Networks Through Feature Decomposition and Memorization
• JaxSGMC: Modular stochastic gradient MCMC in JAX
• DeepSSM: an emulator of gravitational wave spectra from sound waves during cosmological first-order phase transitions
• Hierarchical Bayesian estimation of motor-evoked potential recruitment curves yields accurate and robust estimates
• The sociospatial factors of death: Analyzing effects of geospatially-distributed variables in a Bayesian mortality model for Hong Kong
• Mixed Hamiltonian Monte Carlo for Mixed Discrete and Continuous Variables
• BoGraph: Structured Bayesian Optimization From Logs for Expensive Systems with Many Parameters
• Probabilistic Answer Set Programming with Discrete and Continuous Random Variables
• Evaluating probabilistic programming languages for simulating quantum correlations
• Constraining the primordial power spectrum using a differentiable likelihood
• On decision-theoretic model assessment for structural deterioration monitoring
• D3p -- A Python Package for Differentially-Private Probabilistic Programming
• Gaussian Embedding of Temporal Networks
• Automating Model Comparison in Factor Graphs
• Estimation of Counterfactual Interventions under Uncertainties
• Multi-Fidelity Reinforcement Learning with Gaussian Processes
• Hamiltonian Monte Carlo for Regression with High-Dimensional Categorical Data
• Tuning Random Generators: Property-Based Testing as Probabilistic Programming
• Integrating Markov processes with structural causal modeling enables counterfactual inference in complex systems
• Variationally Inferred Sampling Through a Refined Bound for Probabilistic Programs
• Multi-Turn Beam Search for Neural Dialogue Modeling
• Gibbs Sampling using Anti-correlation Gaussian Data Augmentation, with Applications to L1-ball-type Models
• EffCNet: An Efficient CondenseNet for Image Classification on NXP BlueBox
• A Bayesian Optimization through Sequential Monte Carlo and Statistical Physics-Inspired Techniques
• Identifying latent distances with Finslerian geometry