Knowledge Management - tacit knowledge
The goal of this research is to provide a means by which the tacit knowledge
in an e-mail archive can be browsed in an organized manner and linked
directly to the explicit knowledge found in PubMed. The Pediatric Pain
Mailing List (PPML) is an international Internet-based forum for informal
discussion of any topic related to pain in children. There are now over
seven hundred members, including clinicians, researchers and patients from
at least forty countries on six continents. Currently, the archive contains
more than ten thousand messages.
Automatic Extraction of Metadata
This thesis will investigate the automatic extraction of metadata from Web
documents. Metadata is "information about information", i.e., it describes
the contents of the document or web page. In particular, this thesis will
concentrate on medical related documents and identifying semantic
themes and extracting the metadata based on these themes. This thesis
will use XML and RDF for description of the metadata.
The Use of Genre for Filtering of Information
Genre is normally discussed in the context of rhetoric. It has content and
form and this allows us to identify similar types of documents among a host
of different types. The user, having encountered one instance of a genre,
knows what to expect from any other member of the genre. Examples of genres
are newspapers, mystery novels, cowboy movies, etc. Cybergenres are genres
that have content, form and functionality where functionality is afforded
through the web. Examples include home pages, games, FAQ's, etc. As
different genres migrate to the web they evolve and may develop into new
genres. This research will investigate how to identify the different
cybergenre on the web and how to incorporate cybergenre in the search
process and in organizing the results of a web search.
Context-Aware Information Retrieval
A recent workshop has identified context-aware information retrieval as a particularly important long-term "grand challenge" of information retrieval research that "combine[s] search technologies and knowledge about query and user context into a single framework in order to provide the most "appropriate" answer for a user's information needs."
The goal of this research topic is to develop a model of the context of the information need. The objective is to develop a context model based on a network of classifiers or information agents (mixture of experts) that represent different aspects of the context of the information need. Aspects of the context include long-term and short-term interests of the user, user preferences, whether the user is browsing or searching to satisfy an explicit information need, and if the query is for information retrieval or routing. The nodes of the network both compete and collaborate in evaluating each document for possible retrieval and presentation to the user.
Adaptive Games Based on Models of Behaviour
This work will be in conjunction with the
SAGE Project headed by Dr. Watters.
The thrust of the research will be directed towards games for education in
health that can adapt strategies to respond to behavioural models and user
profiles.