Resources and Prototyping about eXplainable Artificial Inteligence (XAI)
A github repository collecting Interpretability papers as issues.
Own collection of papers in the State-of-the-art folder
The data we use in the experiments are available in this Dropbox folder.
A document summarizing current XAI OS Tools (January 2020): [to be added]
Paper in State-of-the-art folder
Introduciton to LIME blog
Open-source Python (BSD-2) LIME software
Papers in State-of-the-art folder
Interactive LRP demos (public page)
Interactive LRP demos (research group page)
LRP Toolbox (BSD-2), MATLAB, Python and Caffe
New one keras-oriented is coming up!
LRP tutorial (orginial)
LRP tutorial (redone by me)
Paper in State-of-the-art-folder and especially the supplementary material (much longer than the paper)
Paper in State-of-the-art-folder
Paper in State-of-the-art-folder
Paper in State-of-the-art-folder
git repo in Software
Building blocks for interpretability Distill article
March, April 2018: simple experiment to find out the informativeness of LRP heatmaps on a squares vs triangles classification via CNN task
Later in 2018: Testing model 'adaptor' idea, pilot- MATLAB NNToolbox <-> LRP Toolbox ?
May, June 2018: Simple LRP experiement with Triangles and Squares dataset.
October 2018: Code and small experiemtns for the IEEE poster.
April 2020: SImple experiment with counting shapes (in Software/cnn-count)