ROAD-R: The Autonomous Driving Dataset with Logical Requirements
This addresses safety and reliability issues in autonomous driving by providing a benchmark for requirement-compliant models, though it is incremental as it builds on existing datasets and methods.
The paper tackles the problem of neural networks in autonomous driving violating logical constraints by introducing ROAD-R, the first dataset with formally specified requirements, and shows that using these constraints improves model performance and guarantees compliance.
Neural networks have proven to be very powerful at computer vision tasks. However, they often exhibit unexpected behaviours, violating known requirements expressing background knowledge. This calls for models (i) able to learn from the requirements, and (ii) guaranteed to be compliant with the requirements themselves. Unfortunately, the development of such models is hampered by the lack of datasets equipped with formally specified requirements. In this paper, we introduce the ROad event Awareness Dataset with logical Requirements (ROAD-R), the first publicly available dataset for autonomous driving with requirements expressed as logical constraints. Given ROAD-R, we show that current state-of-the-art models often violate its logical constraints, and that it is possible to exploit them to create models that (i) have a better performance, and (ii) are guaranteed to be compliant with the requirements themselves.