Tuesday, September 23, 2014

DREAM

At the beginning of this month, I formally started work on the EU FP7 DREAM project (although the project itself started in April this year). Given that ALIZ-E finished at the end of August, this fit very well for me personally, as it means that I am able to stay in Plymouth. It is coordinated by the University of Skovde (Sweden), with Plymouth (PI Tony Belpaeme) as one of seven partners who between us cover a wide range of expertise. There are two standard robot platforms that will be used as part of the project: the Aldebaran Nao (with which I have plenty of experience in ALIZ-E), and Probo (a green soft-bodied and trunked robot developed by VUB), although the Nao will be the primary focus of development.


(From the nice new flashy project website) The DREAM project...
...will deliver the next generation robot-enhanced therapy (RET). It develops clinically relevant interactive capacities for social robots that can operate autonomously for limited periods under the supervision of a psychotherapist. DREAM will also provide policy guidelines to govern ethically compliant deployment of supervised autonomy RET. The core of the DREAM RET robot is its cognitive model which interprets sensory data (body movement and emotion appearance cues), uses these perceptions to assess the child’s behaviour by learning to map them to therapist-specific behavioural classes, and then learns to map these child behaviours to appropriate robot actions as specified by the therapists.
My work on this will be on the (robot) cognitive and behavioural aspects of this goal. While this is a slight departure from my memory-centred work in ALIZ-E, it remains in the context of child-robot interaction, retains a focus on application-focused development (though for autistic children rather than diabetic children), and maintains an emphasis on achieving autonomous operation (although in the context of supervised interactions). There is an exciting programme of aims and goals in place, and a very good group of partner institutions, so I'm looking forward to it!

Sunday, April 06, 2014

AISB Symposium on New Frontiers in Human-Robot Interaction last week

At the end of last week, I was at the AISB 2014 symposium on New Frontiers in HRI, held at Goldsmiths College, London. Held over two days (well, not quite full days...), it was modestly sized (around 30 on the first day, fewer on the second), but with plenty of time given over to discussion, it was a good size to have the majority of people engage in talking and discussing with one another. The range of topics was quite eclectic (within the HRI theme of course), but there were enough people with wide enough interests/experience to keep the discussion going to greater extent than I was expecting - sometimes, I find that scheduled discussions fall a bit flat with a few too many awkward silences, but that wasn't the case here.

I gave a little overview of my work on applying a memory-centred cognition perspective to social human-robot interaction. I don't think I did particularly well at summarising the particular approach I take, but reckon I at least managed to get across some of the broad ideas involved. Seeing as my starting point is relatively uncontroversial (my assertion that memory is in some way involved in all aspects of human-robot interaction - indeed any social interaction), this was what I hoped. In any case, I tried to make a clear distinction between the the theoretical framework I was proposing, and the actual computation model I have developed to start to explore this question. The reason for this is that I know there are flaws in my implementation (as a model, this is in principle inevitable), but I didn't want people to dismiss the underlying idea as a consequence of this. The general panel discussion afterwards was mostly centred on the role of emotion in HRI (which was a fascinating discussion), but there was a little on these memory-centred ideas too, in which I tried to reinforce the theory rather than my implementation of it (who knows how successfully!).

Elsewhere in the symposium, the one general AISB keynote lecture I went to was given by Humberto Maturana. While it was given over skype (bit too far for him to come over in person...), it was still great to hear him speak in (virtual) person, as someone whose ideas had been such a profound influence on multiple fields of research, and of course my own. It was quite a wide-ranging talk, encompassing many different fields, but all tied together with the central theme of autopoiesis for which he is known. He structured this around four man questions: (1) what kind of systems are robots; (2) what kind of systems are living systems; (3) how does the nervous system operate; and (4) how do we operate as self-conscious beings? The result was an eclectic mix of topics (from life and death, to language, to ethics, and more besides), but which still maintained a coherency that I could follow (as far as I could understand at least!).

Humberto Maturana giving one of the AISB 2014 plenary talks over Skype.

Finally, just a quick note on the workshop I organised last month at the HRI conference with Greg Trafton on Cognitive Architectures for HRI. It was very well subscribed to (over 30 people for the session, which compared very favourably to other workshops running simultaneously), which was more than we were expecting. The invited speakers (Yiannis Demiris, Matthias Scheutz, and Angelo Cangelosi) provided a very good structure, covering a range of topics and perspectives that (according to the various comments from people afterwards) were engaging and indeed appreciated in terms of providing a different take that some were familiar with. Indeed, this also proved to be one of the drawbacks to the workshop: given the half day (and a few little technical issues...) we didn't have the time for discussions of any great length. I take it as a good sign though that people were dissapointed by this, and is certainly something that could be rectified if we were to organise another one.

Friday, December 20, 2013

HRI 2014 Workshop on Cognitive Architectures for Human-Robot Interaction

I am co-organising a half-day workshop at the 9th ACM/IEEE International Conference on Human-Robot Interaction next year (HRI'14), that will be held on the 3rd of March 2014 in Bielefeld, Germany. If you have any interest in this topic, or even would just like to find out more, please consider joining us! We have intended that this will be an inclusive event, with a high discussion content, and an emphasis on dissemination of ideas that will hopefully influence ongoing (social) Human-Robot Interaction research.

I've been interested in cognitive systems, cognitive robotics and cognitive architectures for a while now as my interest (and subsequent research) is in exploring general principles of cognition/intelligence, both for understanding natural systems and for the development of 'better' robotic systems (to test theories and accomplish tasks). I think that Human-Robot Interaction provides a fascinating context to explore cognitive architectures, as it provides a very different set of challenges to theory and implementation than have typically been considered. Hence the workshop!


*******************************************************************
CALL FOR WORKSHOP SUBMISSIONS AND PARTICIPATION

HRI 2014 Workshop on Cognitive Architectures for Human-Robot Interaction

Monday 3rd March, 2014 (Bielefeld, Germany)

http://www.tech.plym.ac.uk/socce/staff/paulbaxter/cogarch4hri/

*******************************************************************

IMPORTANT DATES
***************
** Submission deadline: Friday 10th January, 2014
** Notification of acceptance: Monday 20th January, 2014
** Final (accepted) submission: Friday 7th February, 2014
** Workshop: Monday 3rd March, 2014 (half day)

DESCRIPTION
***********
Cognitive Architectures are constructs (encompassing both theory and models) that seek to account for cognition (over multiple timescales) using a set of domain-general structures and mechanisms. Typically inspired by human cognition, the emphasis is on deriving a set of principles of operation not constrained to a specific task or context. This therefore presents a holistic perspective: it forces the system designer to initially take a step back from diving into computational mechanisms and consider what sort of functionality needs to be present, and how this relates to other cognitive competencies. Thus the very process of applying such an approach to HRI may yield benefits, such as the integration of evidence from the human sciences in a principled manner, the facilitation of comparison of different systems (abstracting away from specific computational algorithms), and as a more principled manner to verify and refine the resultant autonomous systems.

For HRI, such an approach to building autonomous systems based on Cognitive Architecture, 'cognitive integration', would emphasise first those aspects of behaviour that are common across domains, before applying these to specific interaction contexts for evaluation. Furthermore, given inspiration from human cognition, it can also inherently take into account the behaviour of the humans with which the system should interact, with the intricacies and sub-optimality that this entails.

To date, there have been relatively few efforts to apply such ideas to the context of HRI in a structured manner. The aim of this workshop is therefore to provide a forum to discuss the reasons and potential for the application of Cognitive Architectures to autonomous HRI systems. It is expected that by attending this workshop and engaging in the discussions, participants will gain further insight into how a consideration of Cognitive Architectures complements the development of autonomous social robots, and contribute to the cross-fertilization of ideas in this exciting area.

SUBMISSION AND PARTICIPATION
****************************
Contributions are sought from all who are interested in participation. A light-touch review process will be applied to check for relevance - the emphasis of the workshop is on inclusion, discussion and dissemination. Prior to the workshop, the organizers will integrate these into a list a perspectives that will form the basis for the discussions.

Please prepare a 2-page position paper on your research-informed perspective on cognitive architectures for human-robot interaction (particularly social). The HRI template should be used for this submission (ACM SIG Proceedings). Submissions should be sent to: paul.baxter(a)plymouth.ac.uk All accepted position papers will be archived on the workshop website.

ORGANISERS AND CONTACT
**********************
** Paul Baxter (Plymouth University, U.K.) and Greg Trafton (Naval Research Laboratory, USA)
** Email: paul.baxter(a)plymouth.ac.uk