TAGI

Tractable Approximate Gaussian Inference for Bayesian Neural Network

This repository contain the code for performing the regression and classification benchmarks using the TAGI method for Bayesian Networks.

Go see our YouTube channel where you can find more information.

Reference

Tractable Approximate Gaussian Inference for Bayesian Neural Networks<br/>Goulet, J.-A., Nguyen, L.H., and Amiri, S.<br/>Journal of Machine Learning Research, 2021, 20-1009, Volume 22, Number 251, pp. 1-23. , [JMLR]

Installation

These instructions will get you a copy of the project up and running on your local machine for direct use, testing and development purposes.

Prerequisites

Matlab (version 2016a or higher) installed on Mac OSX or Windows

The Matlab Statistics and Machine Learning Toolbox is required.

Installing

1. Extract the ZIP file (or clone the git repository) in a folder you will be working from.
2. Add the TAGI/ folder and all the sub folders to your path in Matlab : e.g.
   • using the "Set Path" dialog in Matlab, or
   • by running the addpath function from the Matlab command window while adding all sub-folders

Getting started

1. Set the working directory to the folder corresponding to dataset you want to run, e.g. ../TAGI/ToyExample
2. run the .m function having the filename corresponding to the dataset, e.g. ToyExample_1D.m

Built With

• Matlab - Coding

Authors

• Luong Ha Nguyen - Main Development - webpage
• James A-Goulet - Initial code and development - webpage

Acknowledgments

The developpement of this code was financially supported by research grants from Hydro-Quebec, and the Natural Sciences and Engineering Research Council of Canada (NSERC)