DistributionsAD.jl

Automatic differentiation of Distributions using Tracker, Zygote, ForwardDiff and ReverseDiff
Popularity
142 Stars
Updated Last
11 Months Ago
Started In
September 2019

DistributionsAD.jl

CI

This package defines the necessary functions to enable automatic differentiation (AD) of the logpdf function from Distributions.jl using the packages Tracker.jl, Zygote.jl, ForwardDiff.jl and ReverseDiff.jl. The goal of this package is to make the output of logpdf differentiable wrt all continuous parameters of a distribution as well as the random variable in the case of continuous distributions.

AD of logpdf is fully supported and tested for the following distributions wrt all combinations of continuous variables (distribution parameters and/or the random variable) and using all defined distribution constructors:

  • Univariate discrete
    • Bernoulli
    • BetaBinomial
    • Binomial
    • Categorical
    • Geometric
    • NegativeBinomial
    • Poisson
    • PoissonBinomial
    • Skellam
  • Univariate continuous
    • Arcsine
    • Beta
    • BetaPrime
    • Biweight
    • Cauchy
    • Chi
    • Chisq
    • Cosine
    • Distributions.AffineDistribution
    • Epanechnikov
    • Erlang
    • Exponential
    • FDist
    • Frechet
    • Gamma
    • GeneralizedExtremeValue
    • GeneralizedPareto
    • Gumbel
    • InverseGamma
    • InverseGaussian
    • Kolmogorov
    • Laplace
    • Levy
    • Logistic
    • LogitNormal
    • LogNormal
    • Normal
    • NormalCanon
    • NormalInverseGaussian
    • Pareto
    • PGeneralizedGaussian
    • Rayleigh
    • Semicircle
    • SymTriangularDist
    • TDist
    • TriangularDist
    • Triweight
    • Uniform
    • Weibull
  • Multivariate continuous
    • MvLogNormal
    • MvNormal
  • Matrix-variate continuous
    • MatrixBeta
    • Wishart
    • InverseWishart

Get Involved

A number of distributions are still either broken or not fully supported for various reasons. See this issue. If you can fix any of the broken ones, a PR is welcome!