Reason induced visual attention for explainable autonomous driving

TL;DR

This paper introduces an explainable deep learning framework for autonomous driving that generates textual descriptions and visual attention maps, improving interpretability and diagnosing model behavior in safety-critical scenarios.

Contribution

It presents a novel framework that jointly models visual input and natural language to induce visual attention, enhancing interpretability of autonomous driving decisions.

Findings

Strong explainability through attention focus on relevant features

Attention maps provide meaningful explanations of model behavior

Identifies weaknesses and potential improvements in the model

Abstract

Deep learning (DL) based computer vision (CV) models are generally considered as black boxes due to poor interpretability. This limitation impedes efficient diagnoses or predictions of system failure, thereby precluding the widespread deployment of DLCV models in safety-critical tasks such as autonomous driving. This study is motivated by the need to enhance the interpretability of DL model in autonomous driving and therefore proposes an explainable DL-based framework that generates textual descriptions of the driving environment and makes appropriate decisions based on the generated descriptions. The proposed framework imitates the learning process of human drivers by jointly modeling the visual input (images) and natural language, while using the language to induce the visual attention in the image. The results indicate strong explainability of autonomous driving decisions obtained by focusing on relevant features from visual inputs. Furthermore, the output attention maps enhance the interpretability of the model not only by providing meaningful explanation to the model behavior but also by identifying the weakness of and potential improvement directions for the model.