INVITED SPEAKER

Dr. Hongmei (Mary) He (SIEEE, FHEA) obtained her PhD in Computer Science and MSc in Multimedia and Internet Computing from Loughborough University, UK in 2006 and 2003, respectively. Before coming to the UK, she was a senior embedded system engineer at Motorola Design House, China. Currently, she is a Professor in Future Robotics, Engineering and Transport systems at the University of Salford. Dr He had rich postdoctoral research experience for several universities, such as Loughborough University, University of Bristol, Ulster University and University of Kent, and she has completed many projects sponsored by EPSRC, MoD, FP7, Leverhmul Trust, Innovate UK as well as International grants. Her expertise in AI has been explored in a wide range of applications, such as Cognitive Robotics, Cognitive Cyber Security, Data/Sensor Fusion, and Safety & Security of autonomous systems, etc. She is a productive researcher and has published many papers in peer-reviewed journals, national/international conferences, and book chapters.

Dr Hongmei He is a full member in EPSRC peer-review college and EU H2020 ICT Robotics Programme review panel member in 2016 and 2019. She is the secretary of IEEE UK & Ireland RAS Chapter, vice chair of Adaptive & Dynamic Programming and Reinforcement Learning Technic Committee in IEEE Computational Intelligence Society, and working group member of IEEE Technical Ethics (P7000) standard.

Title: The Challenges and Opportunities of Human-Centered AI for Trustworthy Robots and Autonomous Systems

Abstract:

The trustworthiness of robots and autonomous systems (RAS) is at the centre of many research agendas on AI driven autonomous systems. This research systematically investigates for the first time the key aspects of human-centred AI (HAI) for trustworthy RAS in terms of safety, security, human-machine interaction, system health and ethics by identifying the challenges in implementing trustworthy autonomous systems with respect to the five key facets and exploring the role of AI in relation to the five facets of trustworthy RAS. It also presents a new acceptance model for RAS as a framework for human-centred AI requirements, promoting machine intelligence that augments human capabilities and places humans at the centre to achieve trustworthy RAS by design.