Opponent: Rafael Peñaloza Nyssen, Associate Professor, University of Milano-Bicocca

Abstract Knowledge bases are used in several industries to efficiently represent data but they sometimes contain errors. Debugging knowledge bases is, therefore, a vitally important process. It requires explanations of entailments in description logic which are used for answering knowledge base queries. Such explanations can be generated by axiom pinpointing, which is a technique to find justifications: minimal sets of axioms that entail a certain conclusion. Typically, a large set of justifications is found. It is difficult to select the one(s) which requires the least cognitive effort to understand. Four contributions are made with this thesis to solve this problem. First, the concept of relative cognitive complexity is defined. Second, the ACT-R model SHARP is created. It simulates the process of a human deciding the consistency of so-called ABoxes in the description logic \(\mathcal{ALE}\), the definition of which some justifications satisfy. SHARP is used for modelling cognitive effort and captures the relative cognitive complexity of \(\mathcal{ALE}\), ABoxes. Third, an experiment is performed to test the predictions on cognitive behaviour based on SHARP’s simulation results. The model performs quite well on the relative cognitive complexity. Fourth, three surrogate modelling techniques were tested: Random Forests (RF), Support Vector Regression (SVR) and Symbolic Regression (SR). The three techniques achieve similar performance, but SR achieves the lowest computation times.

Examination committee

  • Associate Professor Nina Gierasimczuk, Technical University of Denmark
  • Professor Fredrik Stjernberg, Linköpings universitet
  • Senior Research Fellow Paul Mulholland, The Open University
  • Professor Christine Howes, Göteborgs universitet (substitute)