Dependable AI in Safe Autonomous Systems

This project aims to develop machine learning (ML) models with a robust nature for dependable systems, focusing on the dataset side of the problem including novel methods of data augmentation like neural augmentation. While data-driven development methods hold promise for accurate models in perception functions, most lack a holistic view for implementation in dependable systems. By emphasizing the dataset, this project aims to meet emerging certification requirements and set the basis for a safety-conscious ML system, with outputs including a methodology to refine and iterate such systems.

Project manager at MDU: Masoud Daneshtalab