InSecTT will be present to the International Joint Conference on Neural Networks (IJCNN’21) with two publications related to the transferability property of adversarial examples. (arxiv 2104.12679 and 2004.04919). Among its core objectives, InSecTT aims at improving the security of embedded A.I. models against threats focused on striking their integrity.
The whitepaper written by Peter Moertl & Nikolai Ebinger provides processes and guidelines to develop AI that is trustworthy from