The Layer-wise Relevance Propagation (LRP) algorithm explains a classifer's prediction specific to a given data point by attributing relevance scores to important components of the input by using the topology of the learned model itself.
The LRP Toolbox provides simple and accessible stand-alone implementations of LRP for artificial neural networks supporting Matlab and python. The Toolbox realizes LRP functionality for the Caffe Deep Learning Framework as an extension of Caffe source code published in 10/2015.
The implementations for Matlab and python shall serve as a playing field to familiarize oneself with the LRP algorithm and are implemented with readability and transparency in mind. Models and data can be imported and exported using raw text formats, Matlab's .mat files and the .npy format for python/numpy.
For whichever language / purpose you wish to make use of this toolbox download the appropriate sub-package (python, matlab, caffe-master-lrp -- or do a full clone of the project) and then just run the installation script for your implementation of choice, e.g.
git clone https://github.com/sebastian-lapuschkin/lrp_toolbox/
cd lrp_toolbox/$yourChoice
bash install.sh
Make sure to at least skim through the installation scripts! For more details and instructions please refer to the manual.
When using (any part) of this toolbox, please cite our paper
@article{JMLR:v17:15-618,
author = {Sebastian Lapuschkin and Alexander Binder and Gr{{\'e}}goire Montavon and Klaus-Robert M{{{\"u}}}ller and Wojciech Samek},
title = {The LRP Toolbox for Artificial Neural Networks},
journal = {Journal of Machine Learning Research},
year = {2016},
volume = {17},
number = {114},
pages = {1-5},
url = {http://jmlr.org/papers/v17/15-618.html}
}
For further research and projects involving LRP, visit heatmapping.org