Knowledge acquisition from corresponding domain knowledge transformations

  • Authors:

  • Michael Armella;Isaí Michel Lombera;Stuart H. Rubin;Shu-Ching Chen;Gordon Lee

  • Affiliations:

  • Distributed Multimedia Information Systems Laboratory, School of Computing and Information Sciences, Florida International University, Miami, FL;Dept. of Electrical & Computer Engineering, San Diego State University, San Diego, CA;SPAWAR Systems Center, San Diego, CA;Distributed Multimedia Information Systems Laboratory, School of Computing and Information Sciences, Florida International University, Miami, FL;Dept. of Electrical & Computer Engineering, San Diego State University, San Diego, CA

  • Venue:
  • IRI'09 Proceedings of the 10th IEEE international conference on Information Reuse & Integration
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The capability to efficiently retrieve knowledge in response to specific user queries offers the potential to create decision support systems of unprecedented utility, i.e., systems which can accelerate the learning process. This paper presents such an architecture, the Type 2 Knowledge Amplification by Structured Expert Randomization (T2K) system. This system differs from traditional expert systems in the way knowledge rules are matched with queries. The T2K has the ability to acquire knowledge from corresponding domains to answer queries from domains in which the system has less knowledge. This system also solves the word mismatch problem by modifying queries using word substitutions. This is done through creative transformations and optimizations of knowledge rule antecedents and consequents. By pairing rules with identical antecedents or consequents, we are able to induce new rules from existing knowledge without explicit elicitation from the user. The technique presented in this paper attempts to transform both the rules in the knowledge base as well as the query in order to find a matching action for a specified query.