with Professor Tony Belpaeme
When: Tuesday 14 September 2021 12.00 pm – 13.00 pm
Title: Towards autonomous social behaviour in robots
People readily project social competence onto simple robots, and while this is enough to briefly evoke the impression that the robot is socially intelligent, it just is not enough to maintain an interesting interaction over a longer period of time. As autonomous social interactions with robots still are fraught with technical challenges, most demanding interactions still very much rely on tele-operation. This talk will explore how we can move away from tele-operation using interactive machine learning. The SPARC (supervised progressively autonomous robot competencies) method lets the robot monitor commands issued by a tele-operator and gradually takes over, without compromising the control of the operator or the risk of costly exploration steps in more traditional machine learning.
Senft, E., Lemaignan, S., Baxter, P. E., Bartlett, M., & Belpaeme, T. (2019). Teaching robots social autonomy from in situ human guidance. Science Robotics, 4(35). https://robotics.sciencemag.org/content/4/35/eaat1186
Tony Belpaeme is Professor at Ghent University and Professor in Robotics and Cognitive Systems at the University of Plymouth, UK. He received his PhD in Computer Science from the Vrije Universiteit Brussel (VUB) and currently leads a team studying cognitive robotics and human-robot interaction. Recently, he coordinated the H2020 L2TOR project, studying how robots can be used to support children with learning a second language, and coordinated the FP7 ALIZ-E project, which studied long-term human-robot interaction and its use in paediatric applications. He worked on the FP7 DREAM project, studying how robots can be used to support Autism Spectrum Disorder therapy. Starting from the premise that intelligence is rooted in social interaction, Belpaeme and his research team try to further the science and technology behind artificial intelligence and social human-robot interaction. This results in a spectrum of results, from theoretical insights to practical applications.