Verified artificial intelligence (AI) is the goal of designing AI-based systems that that have strong, ideally provable, assurances of correctness with respect to mathematically-specified requirements. This paper considers Verified AI from a formal methods perspective. We describe five challenges for achieving Verified AI, and five corresponding principles for addressing these challenges.
@article{arxiv.1606.08514,
title = {Towards Verified Artificial Intelligence},
author = {Sanjit A. Seshia and Dorsa Sadigh and S. Shankar Sastry},
journal= {arXiv preprint arXiv:1606.08514},
year = {2020}
}