Academic
Publications
Context-sensitive refinements for stochastic optimisation algorithms in inductive logic programming

Context-sensitive refinements for stochastic optimisation algorithms in inductive logic programming,10.1007/s10462-010-9181-y,Artificial Intelligence

Context-sensitive refinements for stochastic optimisation algorithms in inductive logic programming  
BibTex | RIS | RefWorks Download
We describe a new approach to the application of stochastic search in Inductive Logic Programming (ILP). Unlike traditional approaches we do not focus directly on evolving logical concepts but our refinement-based approach uses the stochastic optimization process to iteratively adapt the initial working concept. Utilization of context-sensitive concept refinements (adaptations) helps the search operations to produce mostly syntactically correct concepts. It also enables using available background knowledge both for efficiently restricting the search space and for directing the search. Thereby, the search is more flexible, less problem-specific and the framework can be easily used with any stochastic search algorithm within ILP domain. Experimental results on several data sets verify the usefulness of this approach.
Journal: Artificial Intelligence Review - AIR , vol. 35, no. 1, pp. 19-36, 2011
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.