All Times EDT
Keywords: Explainable AI, AI in regulatory submissions, reproducibility
Two presentations, provided by one presenter from regulatory and by one presenter from industry will serve as a starting point. The panel will consist of the two presenters (Dr. Rantou from FDA and Dr. Köchert from Bayer) of the session and additional representatives from both regulatory and industry (confirmed: Dr. Herbert Pang from Genentech, Dr. Yijie Zhou from Vertex Pharmaceuticals). The panel discussion will start with a rejoinder provided by a representer from academia (confirmed Prof. Tim Friede) and will focus on the general topics of application of AI/ML in a regulated setting circulating on the following general themes: • Which insights generated using AI/ML-methods can serve as supportive evidence in a traditional NDA? • How can reproducibility of AI/ML-methods be assured? • Are AI/ML methods mature enough to use them for guiding treatment decisions? • Are the results typically obtained “explainable enough” or would a standard be necessary? There will also be ample opportunity for the audience to ask questions triggered by the two presentations that preceded the panel discussion