My robots are rich
I am currently developing a new online tool to study social interaction in robotics, in a playful and... COVID-friendly way! I plan to have it both as a tool to teach robotics (incl. ROS) & human-robot interaction, but also as a research platform to conduct more immersive and interactive online studies.
First release expected early 2021. Stay tuned!
Really excited to be starting soon (and despite the COVID pandemic!) an ambitious study in one of Bristol's schools for children on the autism spectrum. The project has already generated a fair bit of excitment, and I am both happy and lucky to work on it with the great UWE colleagues Nigel Newbutt and Louis Rice.
First focus groups with teachers and pupils slated for January 2021, with the study (we will leave a social robots for 3 weeks in the school) starting next Spring. (...in the meantime, programming time for me!)
Since this month, I am now officially Associate Professor in Social Robotics and AI.
Big shout-out to Chris Melhuish and Tony Pipe who both tirelessly supported my application!
I have been invite by the Research Executive Agency of the European Commission as a panelist to the Horizon 2020 MSCA cluster event "Artificial Intelligence". I did present some of the key challenges arising in social HRI, and child-robot interaction in particular, in front of about 150 participants, including the unit heads of several of the EU DGs. Policy making in the making!
Very happy to announce that Emmanuel's landmark study on social robots automatically learning challenging social policies using Interactive Reinforcement Learning has been published in Science Robotics (open access)!
The article presents a novel human-in-the-loop machine learning approach to implement social autonomy in a robot, with several deployments in UK public schools. This is a first-in-kind demonstration of learning autonomous action policy in a high dimensional, socially complex, environment.
With the brilliant Katie Winkle, we just completed an epic study: 9 participants, engaging for 3 months in exercise coaching with a socially autonomous robot! Paper to follow (as soon as Katie has finished writing her PhD thesis ;-) Big media coverage as well, with many radios and TV visiting us on campus!
Edit 2020 a first paper has been published at RSS. More to come soon!
Following Maddy's paper last month, Chris's work on dynamic generation (and repair!) of ambiguous language for human-robot interaction has been published in Frontiers in Robotics and AI (open access). Well done!
The article challenges the common understanding that robots should be unambiguous: we show that ambiguity is often desirable for fluid and natural human-robot interactions, and often lead to better performances, by not overloading the human participant with irrelevant informations.
Just published! Madeleine's work on inferring internal states has been published in Frontiers in Robotics and AI (open access).
The paper investigates how partially hidden 'internal states' (like emotions, cooperativeness, etc) can be decoded from simple visible cues, like skeletons. Also demonstrates that social situations can be described along 3 simple dimensions: the imbalance of the interaction, its general valence, and the degree of engagement of the interactors.
Inspiring visit in Hong-Kong, where I was invited as external expert to report on the findings of the Robot4SEN project: 70 robots deployed in 30 schools for children with special needs, over a period of 2 years. Fascinating and very humbling to see all these robots being adopted by pupils and teachers, and blended into the daily teaching!
Warm thanks to our hosts, the HK Vocational Training Council for the great organisation.
I gave this month a keynote at the UK Robotics & Autonomous Systems conference on HRI, cognition and 'bigg-ish data'. Slides are online!
I've just been invited by Bilge Mutlu to join the editorial board of the freshly launched Human-Robot Interaction section of the open-access Frontiers in Robotics and AI.
Looking forward receiving your submissions!
Just spent 2 days in Istanbul, taking part to the Symposium on Robots for Language Learning.
I gave a talk on what would it take to go from where we are to a broad, ubiquitus use of robots in the classroom. Slides are available online!
Our paper The PInSoRo dataset: Supporting the data-driven study of child-child and child-robot social dynamics presenting the details of the PInSoRo dataset has just been published in open-access in PLOS ONE. If you are looking for a large (45+ hours) dataset of annotated, naturalistic child-child and child-robot social interaction, have a read, and download the dataset!