A method for training neural networks that are provably robust to adversarial attacks.
Primary LanguagePythonMIT LicenseMIT