Kuper, LindseyKatz, GuyGottschlich, Justin EJulian, KyleBarrett, ClarkKochenderfer, Mykel J2023-05-222023-05-222018-01-012020-12-18https://repository.upenn.edu/handle/20.500.14332/8486The increasing use of deep neural networks for safety-critical applications, such as autonomous driving and flight control, raises concerns about their safety and reliability. Formal verification can address these concerns by guaranteeing that a deep learning system operates as intended, but the state of the art is limited to small systems. In this work-in-progress report we give an overview of our work on mitigating this difficulty, by pursuing two complementary directions: devising scalable verification techniques, and identifying design choices that result in deep learning systems that are more amenable to verification.Presentation