ML Component Systemic Hazard Envelope project, or MLCSHE (pronounced /'mɪlʃ/), is a cooperative coevolutionary search algorithm that automatically identifies the hazard boundary of a ML component in an ML-enabled Autonomous System (MLAS), given a system safety requirement.
This work is done at Nanda Lab, EECS Department, University of Ottawa.
Identifying the Hazard Boundary of ML-enabled Autonomous Systems by Sepehr Sharifi, Donghwan Shin, Lionel C. Briand, and Nathan Aschbacher, arXiv pre-prints, January 2023, DOI:XXXX
The details on setting up MLCSHE are provided in the installation guide.
The instructions for running MLCSHE and baseline search methods is provided in usage guide.
The project currently runs on Pylot case study. More details on the case study, the encodings, accessing and reproducing the results is provided in pylot/README.md.
This project is licensed under the MIT License - see the LICENSE file for details.