Mary-Anne Williams


Mary-Anne Williams FTSE is an Australian academic, Distinguished Research Professor at University of Technology Sydney, Fellow at CODEX at Stanford University, and Director of the UTS Magic Lab.
Professor Williams is a Data Scientist and Behavioural Designer with expertise in Artificial Intelligence, Cognitive Science, Disruptive Technologies, Digital Transformation, Business and Law. She is listed among Robohub's "Top 25 Women in Robotics", and celebrated on the First International Day of Women and Girls in Science.
Williams is a Fellow of the Australian Academy of Technological Sciences and Engineering, and a Fellow of the Australian Computer Society. She is on the Digital Transformation and the AI Preparedness Committees at ATSE.
She has been a speaker at major events including the 2016 World Science Festival., United Nations WSIS Forum on the Impact of AI, Australian Strategic Policy Institute. She shared her views on the impact of AI on Human Rights during a panel at the Australian Human Rights Commission Technology Conference.
Williams works on AI, Robotics and Law. She leads a partnership with the South Western Sydney Local Health District, the Softbank Social Robotics Partnership and the UTS partnership with the Commonwealth Bank in Social Robotics. She discussed the impact of Artificial Intelligence on compassion and human rights with the Dalai Lama in Sydney in June 2018,.
Williams has a PhD in Computer Science and a Master of Laws. She is co-founder of the AI Policy Hub. She continues to lead the UTS RoboCup Team that she established at UTS in 2002. This team were the Top International team in 2004, won the Human-Robot Interface Award in 2017. In 2018 UTS RoboCup Team won the Tour Guide Challenge with the highest score of any team on any test in the history of the Social Robotics League. In 2019 her Research Team won the Social Robotics League at RoboCup 2019. They are the current World Champions in Social Robotics.
Williams is known for her foundational contributions to the field of Decision Making using insights, methods and techniques from belief revision. Belief Revision is a fundamental area in Artificial Intelligence. It provides representations and mechanisms for computers to develop a set of beliefs and to revise them over time as they receive new information. Belief revision plays a critical role in Explainable Artificial Intelligence. It allows AI systems to generate explanations of their behaviour that help human interpret, understand, predict, and importantly trust AI systems.
Over the last three decades she has provided solutions to several open research problems in decision making related to finite representations of beliefs, the iteration of belief revision mechanisms, and the relevance of changes and explanations. She developed the first computational models and anytime algorithms for Belief Revision Operators to be applied to real-world problems. Anytime algorithms have a critically important feature for real-world applications, the more time they have the better there outcomes. Not all algorithms have this feature. For example, venturing down fruitless decision/search tree branches usually means backtracking to a weaker outcome.

Publications

*