Research summary
Software is central to most areas of our lives: from health care to finance, transportation, human resources and city management. It comprises code traditionally written by developers and machine-learning-based models generated by analyzing large volumes of data. Unfortunately, software-related failures, such as incorrect and unexpected behaviours, security breaches and privacy violations, are common in all application areas and all types of software.
Dr. Julia Rubin, Canada Research Chair in Trustworthy Software, is devising new theories and approaches to make software and machine-learning-based systems more trustworthy—meaning more reliable, secure, explainable and fair. By leveraging and contributing to the fields of automated program analysis, formal methods, trustworthy machine learning, and empirical software engineering, Rubin and her group build approaches to identify and eliminate software failures, to improve the quality of the software we use. Their approaches also aim to reduce the costs of software development, making developers and machine-learning analysts more efficient and productive.