SOCIALLY INTELLIGENT AGENTS - THE HUMAN IN THE LOOP


                Submission Deadline 29th of March 2000

AAAI Fall Symposium, 3-5 November, Sea Crest Resort, North Falmouth, MA,
USA

The highly interdisciplinary area of Socially Intelligent Agents has
attracted a number of active researchers who model, design and analyse
agents (software or robotic) which behave socially. Much of this work is
strongly inspired by forms of natural social intelligence characteristic
of
humans. This symposium will address recent technological, methodological
and theoretical developments in the field of Socially Intelligent Agents
(SIA's), as well as discuss social and cultural issues, and limitations
and
problems of Socially Intelligent Agents. A focus will be the issue of
the
'human-in-the-loop'. 

Both agents and humans can have different roles during agent-human
interaction, e.g. as designers, users, observers, assistants,
collaborators, competitors, customers, or friends. The symposium will
concentrate primarily on socially intelligent agents that are either
directly interacting with humans, showing aspects of human-style
intelligence, supporting interaction among humans and/or modelling
explicitly aspects of human social intelligence. The symposium will
focus
on four Key Themes for which considerations of the 'human-in-the-loop'
are
crucial. Interdisciplinary approaches are particularly encouraged. 

The symposium will comprise keynote talks, panel discussions and
individual
paper presentations, addressing one or several of the following Key
Themes: 

1) Connecting to SIA's: architectures and design spaces for SIA's;
innovative user-interfaces, novel environments and new methodologies for
software and robotic agents interacting and collaborating with humans
and
facilitating communication and collaboration between humans; hot
approaches
(emotional, empathic aspects) and cold approaches  (intention and plan
ascription, reasoning etc.); synchronisation in human-agent dialogue;
the
role of embodiment in  human-agent interaction; exploiting
anthropomorphism; believability and degrees of agent complexity 

2) Learning and playing with SIA's: new applications of social agent
technology in rehabilitation and education; SIA's as  instructors,
guides,
teachers, assistants and friends; SIA's which support human creativity
and
imagination; SIA's in living environments (e.g. at school, at home, at
work, on holiday, at meeting points) 

3) Living with SIA's: social agent technology which influences
attitudes/opinions/behaviour; issues of 'social relationships'  between
human and agent e.g. helping, competition and cooperation, autonomy and
control, predictability, deception, manipulation, initiative,
delegation,
responsibility, conflicts 

4) Growing up and evolving with SIA's: social agent technology which
empowers humans, addressing the cognitive and emotional needs of humans;
impact of SIA's on human society and culture; agents adapting to and
supporting cultural diversity; ethical considerations 

Submission Information (Deadline 29th of March 2000):
Potential participants are asked to submit a short paper (3 to 5 pages)
describing their work in this area. Please send submissions via
electronic
mail to Kerstin Dautenhahn at [log in to unmask] The text
can either be submitted in plain Ascii format (preferred), or the
submission can be made available on a Webpage and the URL is sent via
email.  

Organizing committee:
     Elisabeth Andre, DFKI GmbH, Germany 
     Ruth Aylett, University of Salford, UK 
     Cynthia Breazeal, MIT AI Lab, USA 
     Cristiano Castelfranchi, Italian National Research Council, Italy 
     Justine Cassell, MIT Media Lab, USA 
     Kerstin Dautenhahn (Chair), University of Reading, UK 
     Francois Michaud, Universite de Sherbrooke, Canada 
     Fiorella de Rosis, University of Bari, Italy 

For updated information on the symposium see
http://www.cyber.rdg.ac.uk/people/kd/WWW/SIA-2000.html or contact
Kerstin
Dautenhahn ([log in to unmask]).


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%