This repository contains the tools and the data of the paper "DeepHyperion: Exploring the Feature Space of Deep Learning-Based Systems through Illumination Search" by T. Zohdinasab, V. Riccio, A. Gambi and P. Tonella, to be published in the Proceedings of the The ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA 2021).

The package is structured as follows:

• DeepHyperion-MNIST contains the DeepHyperion tool adapted to the handwritten digit classification case study and the instructions on how to use it;
• DeepHyperion-BNG contains the DeepHyperion tool adapted to the self-driving car case study and the instructions on how to use it;
• experiments contains the raw experimental data reported in the paper and the scripts to obtain the data.

Note: each sub-package contains further specific instructions.

If you use our work in your research, or it helps it, or if you simply like it, please cite DeepJanus in your publications. Here is an example BibTeX entry:

The software we developed is distributed under MIT license. See the LICENSE.md file.

For any related question, please contact Tahereh Zohdinasab (tahereh.zohdinasab@usi.ch) or Paolo Tonella (paolo.tonella@usi.ch).