The Gaussian Discriminant Variational Autoencoder (GdVAE) - A Self-Explainable Model with Counterfactual Explanations

Copyright © 2024 Ruhr West University of Applied Sciences, Bottrop, Germany AND e:fs TechHub GmbH, Gaimersheim, Germany.

This Source Code Form is subject to the terms of the Apache License 2.0. If a copy of the APL2 was not distributed with this file, You can obtain one at https://www.apache.org/licenses/LICENSE-2.0.txt.

For more info visit project page.

Architecture | Installation | Example | Citation

Architecture

Installation

Anaconda environment

Begin by creating an empty anaconda environment

conda create -n gdvae

Activate the gdvae environment:

conda activate gdvae

Install pip with

conda install pip

Finally install the required packages with

pip install -r requirements.txt 

Example

CelebA Regularized Latent Space

FFHQ Examples

Multi-class MNIST Counterfactual Examples

Citation

If this code is useful to your work, please cite our paper as follows:

@InProceedings{Haselhoff_2024_ECCV,
  author    = {Haselhoff, Anselm and Trelenberg, Kevin and Küppers, Fabian and Schneider, Jonas},
  title     = {The Gaussian Discriminant Variational Autoencoder (GdVAE): A Self-Explainable Model with Counterfactual Explanations},
  booktitle = {European Conference on Computer Vision (ECCV)},
  year      = {2024}
}