Mini-Class

In this mini-class, Grace Gao, assistant professor, aeronautics and astronautics, outlines how to achieve trust with deep learning for vehicle localization is the focus of this mini-class using the following approaches: utilize deep learning as an additional sensor processing tool; integrate deep learning with model-driven approaches; train the deep learning network to learn the sensing uncertainties; and use the learned sensing uncertainty to better quantify pose estimation uncertainty and provide protection levels for guaranteed localization.

Imitation learning involves an expert and building behavior models. Dorsa Sadih, assistant professor, computer science, takes students through problem setup and how to determine the expert policy from expert data, estimate the reward function and then learning a policy from that. Prof. Sadigh's ILIAD lab is conducting extensive work on learning from other sources of data like preferences or physical feedback.

Chris Gerdes, professor, mechanical engineering, takes us through ethical frameworks for vehicle safety from the course, ME267: Ethics and Equity in Transportation Systems. Prof. Gerdes discusses the connection between ethical principles of philosophy and safety principles used in engineering which provide insight and a path forward to reasonable solutions to very difficult problems.