Attentional Bottleneck: Towards an Interpretable Deep Driving Network
This addresses the problem of interpretability in self-driving AI for developers and regulators, though it is incremental as it builds on existing attention and bottleneck methods.
The paper tackles the lack of transparency in deep neural networks for self-driving cars by proposing the Attentional Bottleneck architecture, which improves interpretability by providing sparse attention maps without sacrificing accuracy, showing slight accuracy improvements when applied to the ChauffeurNet model.
Deep neural networks are a key component of behavior prediction and motion generation for self-driving cars. One of their main drawbacks is a lack of transparency: they should provide easy to interpret rationales for what triggers certain behaviors. We propose an architecture called Attentional Bottleneck with the goal of improving transparency. Our key idea is to combine visual attention, which identifies what aspects of the input the model is using, with an information bottleneck that enables the model to only use aspects of the input which are important. This not only provides sparse and interpretable attention maps (e.g. focusing only on specific vehicles in the scene), but it adds this transparency at no cost to model accuracy. In fact, we find slight improvements in accuracy when applying Attentional Bottleneck to the ChauffeurNet model, whereas we find that the accuracy deteriorates with a traditional visual attention model.