Keep the Faith: Faithful Explanations in Convolutional Neural Networks for Case-Based Reasoning

This repository contains the to the paper "Keep the Faith: Faithful Explanations in Convolutional Neural Networks for Case-Based Reasoning"

If you use this code, please cite the following:

@inproceedings{wolf2024keep,
  title={Keep the Faith: Faithful Explanations in Convolutional Neural Networks for Case-Based Reasoning},
  author={Wolf, Tom Nuno and Bongratz, Fabian and Rickmann, Anne-Marie and P{\"o}lsterl, Sebastian and Wachinger, Christian},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={38},
  number={6},
  pages={5921--5929},
  year={2024}
}

Installation

First, create and activate the conda environment

conda env create --file requirements.yaml
pip install --no-deps -e .
conda activate ktf

Usage

In order to train a model, hydra requires the data_dir variable to be set to the folder which contains the data, e.g. /home/datasets:

python train.py data_dir=/home/datasets

Other config variables, e.g. learning rate, model, etc., can be set by appending them to above command call.

Testing a model is done via the test script, which requires the ckpt_path variable to be set. This variable is the path to the pytorch lightning checkpoint of a trained model, e.g. /home/model/checkpoints/epoch=99-bacc.ckpt:

python test.py data_dir=/home/datasets ckpt_path='/home/model/checkpoints/epoch\=99-bacc.ckp'

Utility functions for explanations are available in explain.py.