TrustGAN: Training safe and trustworthy deep learning models through generative adversarial networks

Hélion du Mas des Bourboux

Deep learning models have been developed for a variety of tasks and are deployed every day to work in real conditions. Some of these tasks are critical and models need to be truste…