Imitation is one of the most important mechanisms whereby knowledge is transferred between agents (biological, computational or robotic autonomous systems). This symposium will focus on key problems in this important interdisciplinary area.
The topic of imitation has emerged in various areas close to AI including cognitive and social sciences, developmental psychology, animal behavior, robotics, programming by demonstration, machine learning and user-interface design.
The importance of imitation has grown increasingly apparent to psychologists, ethologists, philosophers, linguists, cognitive scientists, computer scientists, mathematicians, biologists, anthropologists, and roboticists. Yet the workers in the field of imitation are often unaware of relevant research by others in other disciplines. The study of imitation has lacked a rigorous foundation and no major interdisciplinary publication is available on the subject for workers in AI. The symposium is aimed toward remedying this situation and will comprise invited keynote lectures, peer-reviewed contributed presentations, expert panels and general discussion in the interdisciplinary area of imitation. This will be achieved by bringing together established researchers from different areas and producing a publication which can be used as a standard reference in research and teaching for the AI community and others in this exciting field. A rigorously refereed and edited volume including invited and selected contributed papers will be published by a major scientific publisher.
The areas of interest of the Symposium on Imitation in Animals & Artifacts will include, but are not limited to:
Imitation is on the one hand considered as an efficient mechanism of social learning, and experiments in developmental psychology suggest that infants use imitation to get to know others as persons, perhaps by applying a `like-me' test: `persons are objects which I can imitate and which imitate me'. On the other hand, imitation methods as in programming by demonstration setups in robotics and machine learning have primarily focused on the technological dimensions, while disregarding the more social and developmental functions. Additionally, the split between imitation research in natural sciences and the sciences of the artificial has been difficult to bridge, as we lack a common framework supporting an interdisciplinary approach. Yet, studying imitation for an embodied system inhabiting a non-trivial environment leads one to address all major AI problems from a new perspective: perception-action coupling, body-schemata, learning of sequences of action, recognition and matching of movements, contextualization, reactive and cognitive aspects of behavior, the development of sociality, or the notion of `self', just to mention a few issues.
Imitation involves at least two agents sharing a context, allowing one agent to learn from the other. The exchange of skills, knowledge, and experience between natural agents cannot be achieved by brain-to-brain communication in the same way computers can communicate via the Internet. It is mediated via bodies, the environment, the verbal or non-verbal expression or body language of the `sender', which in return has to be interpreted and integrated in the `recipient's' own understanding and behavioral repertoire. Moreover, as imitation games between babies and parents show, the metaphor of `sender' and `receiver' is deceptive, since the game emerges from the engagement of both agents in the interaction (cf. notions of situated activity and interactive emergence). Thus, learning by imitation and learning to imitate are not just a specific topics in machine learning, but can be seen as a benchmark challenges for successful real-world AI Systems.
The symposium homepage is at http://homepages.feis.herts.ac.uk/~nehaniv/aisb.html
Papers will be selected by anonymous peer review of extended abstracts of not more than 4 A4 pages. A cover page should be supplied listing the Title, and the Author's name and affiliation, but the extended abstract itself should not identify the author. Deadlines are listed in the timetable, below.
| Programme Chairs: | |
|---|---|
| Kerstin
Dautenhahn
Department of Cybernetics University of Reading Whiteknights, PO Box 225 Reading RG6 6AY United Kingdom fax: +44-118-931-6220
|
Chrystopher
Nehaniv
Interactive Systems Engineering Faculty of Engineering & Information Sciences University of Hertfordshire Hatfield Herts AL10 9AB United Kingdom fax: +44-1707-284-303
|
Prof. Martin J. Loomes
Department of Computer Science
University of Hertfordshire, U.K.
Programme Committee:
The following formats are acceptable:
Important Dates:
please conform
with the number of pages we have allocated for you for the
proceedings (see
acceptance email).
Preparation of the
camera-ready hardcopy must be done in accordance to
the
instructions available
HERE.
Note that camera-ready hardcopies (only)
should be sent to Dr. Geraint Wiggins at the address
below, not to the symposium chairs!
Local organizers:
Dr Geraint Wiggins & Dr Helen Pain, School of Artificial Intelligence,
Division of Informatics, University of Edinburgh, 80 South Bridge, Edinburgh
EH1 1HN, Scotland
{geraint,helen}@dai.ed.ac.uk; Tel: +44-131-650 2702; Fax:
+44-131-650 651