Seminar: Machine Learning and Formal Verification
Schedule online course:
90 hours, of which 28 hours are spent in class and 62 hours of self study.
Performing quality assurance for software systems relying on machine learning components, such as Deep Neural Networks (DNNs), is difficult since such components to not have internal structure that can directly and easily explain why some result was computed by the component. In this seminar we will review approaches towards quality assurance for machine learning based systems using formal verification and testing. Applications from various domains, including automated driving, will be presented.
This course takes place as an online course.
Further details on the course will be provided in ILIAS.
Please register in ZEuS for the course, which will register you automatically in ILIAS as well.
This seminar aims at presenting some first approaches towards the development of formal verification technology in order to proof properties of DNNs, or to test whether DNNs satisfy certain properties.
Literature will be announced during the preliminary meeting and on the related web-pages.
Bachelor-level and Master-level course.
The course will be taught in English. All course materials will be in English.
Expected course achievement
30 - 45 min. presentation + 15 min. discussion