PhD Seminar • Artificial Intelligence | Machine Learning • Wise Rule: A Rule-Based Behaviour Planner to Top Carla Leaderboard 1

Friday, November 22, 2024 1:00 pm - 2:00 pm EST (GMT -05:00)

Please note: This PhD seminar will take place in E7 5419 and online.

Frédéric Bouchard, PhD candidate
David R. Cheriton School of Computer Science

Supervisor: Professor Krzysztof Czarnecki

Despite the undoubted success of deep neural networks in reproducing human-like driving behaviour, the necessity of rules in behaviour planning for autonomous driving seems unavoidable. In order to specify or validate the safety, reliability and legal conformance of an autonomous vehicle (AV), there must exist some formally defined specification or imperatives that can be enforced and checked. Moreover, the decisions made by human drivers can often be explained in simple terms, suggesting that discrete driving decisions can be encoded as sets of explainable rules. Given that the driving task is nevertheless very complex, the challenge is how to obtain such rules.


To attend this PhD seminar in person, please go to E7 5419. You can also attend virtually using MS Teams.