October 6th 2020, 9-16h | |
9:00 AM | Welcome Note |
9.15 am | Session 1: Life-Cycle and Interpretability> BSI DI11 (Bonn, Germany): Life-Cycle of AI systems > Wojciech Samek (Fraunhofer HHI, Berlin, Germany): From local to global interpretations of DNNs > Barbara Hammer (Bielefeld University, Germany): Online Learning in the Presence of Drift |
10.45 am | Coffee Break |
11.00 am | Session 2: Attack and Defense> Battista Biggio (University of Cagliari, Italy): Adversarial Machine Learning > Aleksander Madry (MIT, Cambridge, MA, USA): Backdoor Attacks on DNNs > Jan Hendrik Metzen (Bosch Center for Artificial Intelligence, Stuttgart, Germany): Detection of and Defenses against Attacks on DNNs |
12.30 | Lunch |
1.30 pm | Session 3: Auditing and Standardization of AI-Applications> Maximilian Poretschkin (Fraunhofer IAIS, Sankt Augustin, Germany): Overview of AI standardization activities worldwide > Heidy Khlaaf (Adelard LLP, London, GB): Auditingsafety-critical AI systems > Martin Vechev (ETH and LatticeFlow, Zürich, Switzerland): Verification of AI Systems |
3.15 pm | Panel discussionShort summary of sessions, listing of open questions, towards auditability of AI systems - further actions and priorities (roadmap) |
4:00 PM | End of the workshop |