Journal of Machine Learning Research

Special Issue on Fusion of Domain Knowledge with Data for Decision Support

Statistics and machine learning are data-oriented tasks in which domain models are induced from data. The bulk of research in these fields concentrates on inducing models from data archived in computer databases. However, for many problem domains, human expertise forms an essential part of the corpus of knowledge needed to construct models of the domain. The discipline of knowledge engineering has focused on encoding the knowledge of experts in a form that can be encoded into computational models of a domain. At present, knowledge engineering and machine learning remain largely separate disciplines. Yet in many fields of endeavor, substantial human expertise exists alongside data archives. When both data and domain knowledge are available, how can these two resources effectively be combined to construct decision support systems?

The aim of this special issue of the Journal of Machine Learning Research is to allow researchers to communicate their work on integrating domain knowledge with data (knowledge-data fusion; theory revision; theory refinement) to a general machine-learning audience. Emphasis is on sound theoretical frameworks rather than ad hoc approaches. Of particular interest are papers that combine clear theoretical discussion with practical examples, and papers that compare different approaches.

Frameworks for knowledge-data fusion include probabilistic (Bayesian/belief) networks, possibilistic logics and networks, hybrid neuro-fuzzy networks, and inductive logic programming.

Topics of interest include (but are not limited to):

• Practical applications of knowledge-data fusion. What lessons have been learnt from attempts to apply knowledge-data fusion to real-world decision problems?
• How are the various knowledge representation and inference frameworks that permit induction theoretically related to each other?
• What frameworks enable an existing induced model, such as a neural network, to be incorporated into a proposed knowledge-based system? How can different knowledge-data hybrid systems be interfaced with each other?
• How can knowledge-data fusion be applied to temporal data?

Schedule:

• Paper submission deadline: June 1, 2001    PAST
• Authors' notification of decisions: October 4, 2002    PAST
• Final revised papers due: November 30, 2002
• Final papers submitted to the JMLR: January 31, 2003

Guest Editors:

Richard Dybowski (InferSpace), richard@inferspace.com.
Kathryn Blackmond Laskey (George Mason University), klaskey@gmu.edu.
James Myers (TRW), James.W.Myers@trw.com
Simon Parsons (Liverpool University), s.d.parsons@csc.liv.ac.uk