Edward: A library for probabilistic modeling, inference, and criticism

By Dustin Tran, Alp Kucukelbir, Adji Bousso Dieng, Maja R. Rudolph, Dawen Liang, David M. Blei. 2016

306 citations recorded by the source corpus.

Probabilistic modeling is a powerful approach for analyzing empirical information. We describe Edward, a library for probabilistic modeling. Edward's design reflects an iterative process pioneered by George Box: build a model of a phenomenon, make inferences about the model given data, and criticize the model's fit to the data. Edward supports a broad class of probabilistic models, efficient algorithms for inference, and many techniques for model criticism. The library builds on top of TensorFlow to support distributed training and hardware such as GPUs. Edward enables the development of complex probabilistic models and their algorithms at a massive scale.

Subfields: Artificial Intelligence, Probabilistic & Bayesian Methods

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Probabilistic & Bayesian Methods

Citations in the corpus

Foundation Posteriors for Approximate Probabilistic Inference (2022, neutral): “Many probabilistic programming languages (PPLs) have been developed [27, 45, 28, 44, 18, 58, 47, 53, 59, 12, 23, 17, 7, 57] each with inference approaches with strengths and weaknesses.”
An Introduction to Probabilistic Programming (2018, methodology): “, 2017) provides implementations of random variable primitives which can be incorporated into deep generative models; Edward (Tran et al., 2016) provides a modeling and inference environment for defining structured, hierarchical distributions for encoders and decoders.”
Causal Effect Inference with Deep Latent-Variable Models (2017, positive): “For the implementation of our model we used Tensorflow [1] and Edward [52].”
Reliable Uncertainty Estimates in Deep Neural Networks using Noise Contrastive Priors (2018, methodology): “Our implementation uses TensorFlow Probability (Dillon et al., 2017; Tran et al., 2016) and is available at https://github.”
Hierarchical Variational Models (2015, methodology): “In other work, we use Gaussian processes (Tran et al., 2016b).”
Automatic differentiation in machine learning: a survey (2015, background): “Edward (Tran et al., 2016) is a library for deep probabilistic modeling, inference, and criticism (Tran et al.”
Automatic Differentiation Variational Inference (2016, background): “Tran et al. (2016) place a Gaussian process to learn the optimal transformation and prove its expressiveness as a universal approximator.”
Deep and Hierarchical Implicit Models (2017, methodology): “To solve this, we consider ratio estimation.”
Implicit Causal Models for Genome-wide Association Studies (2017, unknown): “Appendix B provides an example implementation in the Edward probabilistic programming language (Tran et al., 2016).”
Machine Learning in Python: Main developments and technology trends in data science, machine learning, and artificial intelligence (2020, methodology): “Notable software in this area includes the PyStan59 wrapper of STAN [223], the Theano-based PyMC3 [224] library, the TensorFlow-based Edward [225] library, and Pomegranate [226], which features a user-friendly Scikit-learn-like API.”
Simple, Distributed, and Accelerated Probabilistic Programming (2018, methodology): “To specify random choices in the program, we use RandomVariables from Edward [49], which has similarly been built on by Zhusuan [42] and Probtorch [34].”
INFERNO: Inference-Aware Neural Optimisation (2018, unknown): “which has already been already approached by automatic differentiation in probabilistic programming frameworks [13].”
Advances in Variational Inference (2017, methodology): “mainly rely on sampling methods, and Edward [178], which supports BBVI as well as Monte Carlo sampling.”
Hinge-Loss Markov Random Fields and Probabilistic Soft Logic (2015, background): “…languages include IBAL (Pfeffer, 2001), BLOG (Milch et al., 2005), Markov logic (Richardson and Domingos, 2006), ProbLog (De Raedt et al., 2007), Church (Goodman et al., 2008), Figaro (Pfeffer, 2009), FACTORIE (McCallum et al., 2009), Anglican (Wood et al., 2014), and Edward (Tran et al., 2016).”
TensorFlow Distributions (2017, background): “External developers have built on it to design probabilistic programming and statistical systems including Edward [47] and Greta [17].”
Deep Probabilistic Programming (2017, background): “…(which we discuss in Section 4), we can embed it into more complicated algorithms, such as with expressive variational approximations (Rezende & Mohamed, 2015; Tran et al., 2016b; Kingma et al., 2016) and alternative objectives (Ranganath et al., 2016a; Li & Turner, 2016; Dieng et al., 2016).”
ProBO: a Framework for Using Probabilistic Programming in Bayesian Optimization (2019, methodology): “We implement other PPs in Edward (Tran et al., 2016), where we use black box variational inference algorithms, and in Stan (Carpenter et al., 2015), where we use the No U-Turn Sampler (Hoffman and Gelman, 2011), a form of Hamiltonian Monte Carlo.”