DM: IJCAI Workshop on Automating the Construction of Case-Based Reasoner

[(by way of Dorothy Firsching )]

IJCAI-99 Workshop on Automating the Construction of Case Based Reasoners Venue: City Conference Center, Stockholm, Sweden 2nd August, 1999

Case based reasoning (CBR) was presented to the field of knowledge-based systems as a solution to the knowledge acquisition bottleneck and brittleness, ails from which rule-based systems were known to suffer.

However, CBR systems also require substantial knowledge acquisition effort (e.g. acquiring cases, case vocabulary, retrieval knowledge, adaptation knowledge). Acquiring this knowledge for CBR systems has traditionally been heavily dependent on the availability of a domain expert.

Today most organisations have large operational data sets that, to varying degrees, model various real-world processes. The question arises: Can knowledge implicitly contained within these databases be harnessed using data mining techniques to reduce the domain expert dependence present in case base development?

Data mining is considered to be one of the ten most important technologies (Gartner Group Inc, January 1998), in terms of the potential impact on industry and wide-ranging applications. While CBR and data mining make similar assumptions of regularity, typicality and consistency, they are largely complementary technologies. Data mining focuses on the process of discovering knowledge while CBR focuses on the management and application of knowledge through representation, retrieval, re-use, revision and retention of case knowledge. With major strides being made in knowledge integration within CBR systems, CBR may be viewed as a general knowledge management and problem solving tool. Knowledge management has, to date, not been central to data mining research. Keeping this in mind, this workshop focuses on how CBR and data mining can support each other. In particular, how data mining can aid the construction on case based reasoners.

Data mining may be used to support the acquisition of knowledge required to construct the case-base as well as perform the four CBR functions of retrieve, reuse, revise and retain in a number of ways. However, to do so, a number of research issues must be addressed. These issues will form the primary focus of the workshop and include (not exhaustively)

Representation

    A database is not necessarily a case base; what characterises a 

case base?

    How do you identify a case?

    To what extent can recent developments within data mining 

contribute to

automated case base construction (authoring)?

    How can case extraction (from, for example, documents, structured 

logs,

the World-Wide Web) be supported?



Retrieval

    How is similarity measured between two cases?

    What is the best structure for individual cases and the case base 

as a

whole?

    Can suitable indexing regimes be elicited from data about the 

domain?



Re-use

    What are the sources for adaptation knowledge?

    Can the discovery of useful adaptation knowledge be automated?



Revise

    How can feedback from the application of  solved cases be used to

provide useful knowledge for future applications of the CBR system?

    How can this feedback be used in future applications of the CBR 

system?



Retention

    How do we keep case knowledge up-to-date i.e. learning phase in 

CBR?




Submission Requirements

Authors should submit original papers no longer than 4 pages formatted

according to IJCAI format. Electronic submissions are encouraged in

postscript or pdf format to ss.anand@ulst.ac.uk on or before the 

submission deadline of 1 April, 1999.



Proceedings

Papers accepted for the workshop will be published as a separate IJCAI

working notes series, made available on the day of the workshop to

attendees. In keeping with the fact that IJCAI encourages the 

production of publications based on the workshops, we will

endeavour to follow up the workshop with a special issue on the

topic in an international journal.



Important Dates

1 April, 1999: Submission deadline

1 May, 1999: Acceptance/Rejection Notifications

24 May, 1999: Camera Ready Papers Deadline

2 August, 1999: Workshop





Workshop Participation

The workshop will be kept small, with a maximum of 40 participants.

Preference will be given to active participants selected on the basis 

of their submitted papers.

According to IJCAI rules, all workshop attendees must register for 

the main conference.



Organising Committee

  Sarabjot Singh Anand

School of Information and Software Engineering,

University of Ulster,

Newtownabbey, County Antrim,

Northern Ireland BT37 0QB

E-mail: ss.anand@ulst.ac.uk

Agnar Aamodt

Department of Computer and Information Science,

Faculty of Physics, Informatics and Mathematics,

Norwegian University of Science and Technology,

N-7034, Trondheim, Norway

E-mail: agnar.aamodt@idi.ntnu.no

David W. Aha

Navy Center for Applied Research in AI,

Naval Research Laboratory, Code 5510,

4555 Overlook Avenue, SW,

Washington, DC 20375-5337, USA

E-mail: aha@aic.nrl.navy.mil







Programme Committee

Klaus-Dieter Althoff, Fraunhofer Institute for Experimental Software

Engineering, Germany

Karl Branting, University of Wyoming

Werner Dubitzky, University of Ulster, Northern Ireland

Mark Keane, Trinity College, Dublin, Republic of Ireland



Hiroaki Kitano, Sony Computer Science Laboratory Inc, Japan

David Leake Indiana University, USA

David Patterson, University of Ulster, Northern Ireland

Enric Plaza, Spanish Scientific Research Council, Barcelona, Spain