Why did the AI make that decision? Towards an explainable artificial intelligence (XAI) for autonomous driving systems
Dong, J., Chen, S., Miralinaghi, M., Chen, T., Li, P. and Labi, S. (2023).
Transportation Research Part C: Emerging Technologies, 156, p.104358.
Summary
This study develops an explainable autonomous-driving model that generates textual descriptions of driving scenes and uses those descriptions to support driving decisions. A Transformer-based multimodal architecture connects visual information with language. The generated explanations are intended to help users understand vehicle actions and help developers diagnose model behavior.
Brief summary of the paper; see the publication record for the full abstract.