Learning explainable policies for self-driving cars from little data (17*)

• Subject Area: Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
• Term: from 2021 to 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 276693517

The goal of this project is to learn explainable, robust and generalizable policies for self-driving cars end-to-end from data. Existing approaches to learning self-driving policies end-to-end are limited with respect to two fundamental aspects: generalization and inter-pretability. In this project, we plan to tackle both aspects by combining ideas from modular approaches, representation learning, recur-rent attention and zero-shot learning to yield an introspective model that generalizes to novel driving situations and behaviours.

• DFG Programme: Collaborative Research Centres
• Subproject of: SFB 1233: Robust Vision - Inference Principles and Neural Mechanisms
• Applicant Institution: Eberhard Karls Universität Tübingen
• Project Heads: Professorin Dr. Zeynep Akata; Professor Dr.-Ing. Andreas Geiger