This framework was used by the Bosch Center for AI while participating in the following program: Guaranteeing AI Robustness Against Deception (GARD)
BCAI ART is a framework that provides core capabilities for training and testing models robust to a number of adversarial attacks. This software is a research prototype, solely developed for and published as part of the GARD program cited above. It will neither be maintained nor monitored in any way.
BCAI ART is open-sourced under the AGPL-3.0 license. See the LICENSE file for details.
For a list of other open source components included in PROJECT-NAME, see the file 3rd-party-licenses.txt.
Arash Norouzzadeh, Leonid Boytsov, Wan-Yi Lin