Code for the paper "Partially Exchangeable Networks and Architectures for Learning Summary Statistics in Approximate Bayesian Computation"
Branch: master
Clone or download
Latest commit 1a07102 Jan 30, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
lunarc first commit Jan 24, 2019
notebooks first commit Jan 24, 2019
src first commit Jan 24, 2019
.gitattributes first commit Jan 24, 2019
.gitignore first commit Jan 24, 2019
LICENSE.md updated readme added license Jan 30, 2019
README.md updated readme Jan 30, 2019

README.md

Code for the paper Partially Exchangeable Networks and Architectures for Learning Summary Statistics in Approximate Bayesian Computation

This repository contains all the code for the pre-print paper Partially Exchangeable Networks and Architectures for Learning Summary Statistics in Approximate Bayesian Computation by S. Wiqvist, P-A. Mattei, U. Picchini and J. Frellsen. https://arxiv.org/abs/1901.10230

N.B.: The results in the pre-print at arXiv v1 are computed using the version of the code at tag preprint_v1.

File structure

/lunarc

/notebooks

  • notebooks used for creating the plots in the paper (and several other plots)

/src

  • source files

/src/abc algorithms

  • code for generic ABC rejection sampling algorithm

/src/alpha stable

  • source files for the alpha-stable distribtion

/src/AR2

  • source files for the AR2 model

/src/g-and-k distribtion

  • source files for the g-and-k distribution

/src/generate training test data

  • generic code to generate training and test data for some model

/src/MA2

  • source files for the MA2 model

/src/nets

  • generic code for the MLP network and generic loss functions

/src/utilities

  • help functions (random number generators, distance functions etc.)

Software

The code is written in Julia 1.0.0 and the framework Knet is used to train the networks. The code can be run on both CPUs and GPUs.

The networks are trained on a Nvidia K80 card by utilizing the GPU nodes at LUNARC (http://www.lunarc.lu.se/).

Packages used:

  • Knet v1.1.0
  • Distributions v0.16.4
  • MultivariateStats v0.6.0 master
  • CSV v0.3.1
  • DataFrames v0.13.1
  • StatsFuns v0.7.0
  • StatsBase v0.25.0

Data

The data used in the paper can be generated from the code.