-------- Original-Nachricht -------- Betreff: [all-prolearn] CFP: UM'05 WORKSHOP on Decentralized, Agent Based and Social Approaches to User Modelling Datum: Wed, 2 Mar 2005 10:11:31 +0100 Von: Peter Dolog dolog@l3s.de An: Prolearn-All all-prolearn@agws.dit.upm.es
* Please, excuse cross-posting! * Deadline is in few days (March 7, 2005) * _____________________________________________________________________
UM'05 WORKSHOP on Decentralized, Agent Based and Social Approaches to User Modelling (DASUM) July 25, 2005, Edinburgh
http://www.l3s.de/~dolog/dasum/
_____________________________________________________________________
MOTIVATION: ____________
Time is ripe to discuss decentralized approaches to user modelling, since decentralized applications are becoming prevalent both in web-based and mobile/ ubiquitous environments. Such applications include personal guides or helpers for navigation or ambient devices, integrated web-sites (e.g. newspapers or magazines), portals (e.g. Yahoo), e-commerce web-sites (e.g. Amazon, e-Bay), or recommender sites (e.g. MovieLens). Emerging classes of such applications are loosely coupled systems.
Both, web-service-based and ubiquitous computing applications can be considered as conglomerates of independent, autonomous services developed by independent parties, which have not been integrated by design, but integrate dynamically at run-time, as the need arises. A frequently used metaphor is a free-market of services where the user is a shopper that buys a larger service composed dynamically by smaller services. For example, e-learning courses assembled dynamically from independently created repositories of learning objects and tailored to the needs of a particular learner. Another area where decentralized and social approaches to user modelling are applied is the enrichment of profile data (obtained from whatever source) by information from completely unrelated sources, for example, demographic and sociographic data. An example for this is the Lifestyle Finder and its method of demographic generalization.
In decentralized settings, each small player (agent, smart sensor, mobile device, learning object, application, web-service) maintains a small user model (UM) or profile, as needed for its own purposes of adaptation. These models are updated by the players sporadically, whenever they interact with users. However, the players can also talk with other players who interact with users and build their own models to exchange user information, and to be more up-to-date. In this way, through communication, the agents leverage the benefits of the efforts done by many modelers. Instead of one central model acting as a sink where the subscribed applications report their user data, or instead of having isolated models for each application, we have a community of adaptive applications sharing user information.
QUESTIONS: __________
There are many questions that arise:
- How do we define a (a decentralized) user model? There is no central notion of a model, but user data fragments dispersed among the various devices, services and agents; the level of this data can be hugely different, from sensory data about user and context in ubiquitous computing systems, usage statistics in recommender systems or web-services, to detailed preference models in negotiation agents. Do we talk of all these as "models" or do we define the model in relation to the purpose for which it will be used?
- How to ensure that different autonomous, independently created applications and services are able to communicate with each other and exchange user data, if each represents its user models in a different way (ontologies, communication protocols and languages). Can user data harvested in one context be useful for adaptation in another context? How to interpret a user model so that it can be used in a new context? Should the "little player" do this computation or special components that know what user data is relevant for what type of adaptation? (architectures for adaptation, standard user modelling tasks / purposes).
- How to represent the knowledge necessary to compute user models on demand? What does this knowledge include? Possibly - what user data is relevant for the purpose, where to obtain it from and how to adapt the application/service. Should representations be centered around the user data, the content, or around the purposes of use? Should they be declarative or procedural? Is any data ever irrelevant? What user data to keep and what not to keep after adaptation? (knowledge representation, reasoning, learning).
- Whom to ask for user information? Who is "trusted"? How to define trust in this case: an agent that is honest, or similar to the client agent in its purpose or criteria/preferences, or both? How to use models of interpersonal relationships and how to interpret user data received from acquaintances? (trust, reputation, gossiping).
- How interpret data obtained by other players? How to resolve conflicts between user model fragments kept by different agents (e.g. using information about adaptation and purpose)? How to protect oneself from deliberate deception? When to provide data and to whom?
- Is decentralized and social user modelling a threat for privacy or is it a solution? Do recent developments in industry about Liberty Alliance, Microsoft Passport II, and Identity management help in a distributed environment?
- What are the candidate applications: where do we expect to see such applications first? Recommender systems? Mobile / ubiquitous computing applications or ambient computing environments? E-learning systems? E-commerce systems? What is common among these application areas?
TOPICS: ________
The list of topics (non exclusive) includes:
- agents modelling other agents, - modelling trust and similarity (affinity) between users/agents, - modelling social relationships, networks and groups, - user modelling within mobile and pervasive computing environments - harvesting user and context information from ubiquitous environments - user/agent clustering techniques, - coalition and community formation based on user affinity and trust, - collaborative planning, - architectures, - commonalities and differences between user modelling approaches on the web and in ubiquitous environments. - applications in ubiquitous, agent-based and decentralized systems such as: - recommender systems, expert finding, - web-services, multi-agent systems, - mobile / ubiquitous / ambient computing systems, - on-line communities, - peer-to-peer systems,
AUDIENCE: _________
The workshop is likely to attract researchers with interests in any combination of the following four areas:
- user modeling, - multi-agent systems, - mobile, ubiquitous, pervasive and ambient systems, - user adaptive web-services, e-learning, recommender systems.
The one-day workshop will be held during the 10th International Conference on User Modeling in Edinburgh, Scotland ( http://gate.ac.uk/conferences/um2005/um05.html). The geographic and time proximity with AIED, IJCAI and AAMAS would allow participants from all these research communities to attend.
SUBMISSIONS: ____________
Two types of contributions are invited:
- Papers describing (ongoing) work addressing one or more questions or topics of the workshop (10 page maximum)
- Position statements regarding one or more of the questions/topis of the workshop (3 page maximum)
Both papers and position statements should be prepared according to the UM05 Instructions for Authors available at: (http://www.springeronline.com/sgw/cda/frontpage/0,11855,5-164-2-72376-0,00. html)
and be sent to
Julita at jiv@cs.usask.ca , and Peter at dolog@l3s.de
DEADLINE: _________
**** March 7, 2005. *****
SCHEDULE: __________
Paper submission: March 7, 2005 Notification of acceptance: April 7, 2005 Camera-ready paper submission: May 15, 2005
Each paper and position statement will be reviewed by at least two reviewers. Accepted contributions will be published in the workshop proceedings and will be available on the Web at least 1 month before the workshop. Depending on the quality of the accepted papers, a post- conference book or special issue journal publication is envisaged.
ORGANIZATION: ______________
Chairs:
Julita Vassileva (co-organizer) Computer Science Department University of Saskatchewan 176 Thorvaldson Bldg., 110 Science Place Saskatoon SK, S7N 5C9, Canada jiv@cs.usask.ca
Peter Dolog (co-organizer) L3S Research Center, University of Hannover, Expo Plaza 1, 30539 Hannover, Germany dolog@l3s.de
Program Committee:
Liliana Ardissono, University of Torino, Italy Lora Aroyo, Eindhoven University of Technology, The Netherlands Mathias Bauer, DFKI, Germany Peter Brusilovsky, University of Pittsburgh, PA, USA Susan Bull, University of Birmingham, UK Keith Cheverst, University of Lancaster, UK Robin Cohen, University of Waterloo, Canada Nadia de Carolis, University of Bari, Italy Boris De Ruyter, NatLab / Philips Research, The Netherlands Ludger van Elst, DFKI, Germany Elena Gaudioso,University for Distance Learning, Spain Piotr Gmytrasiewicz University of Illinois, Chicago, USA Dominik Heckmann, DFKI, Germany Judy Kay, University of Sydney, Australia Alfred Kobsa, University of California at Irvine, USA Antonio Krueger, DFKI, Germany Daniel Kudenko, University of York, UK Gord McCalla, University of Saskatchewan, Canada Daniel Olmedilla, L3S Research Center, University of Hannover, Germany Olayide Olorunleke, University of Saskatchewan, Canada Fiorella de Rosis, University of Bari, Italy Boris De Ruyter, Philips Research, The Netherlands Michael Sintek, DFKI, Germany Amy Soller, Institute for Defense Analyses, USA Thomas Tran, University of Ottawa, Canada Marianne Winslett, University of Illinois at Urbana-Champaign, USA