Qualitative Reasoning about Incomplete Categorization Rules Based on Interpolation and Extrapolation in Conceptual Spaces
Various forms of commonsense reasoning may be used to cope with situations where insufficient knowledge is available for a given purpose. In this paper, we rely on such a strategy to complete sets of symbolic categorization rules, starting from background information about the semantic relationship of different properties and concepts. Our solution is based on Gärdenfors conceptual spaces, which allow us to express semantic relationships with a geometric flavor. In particular, we take the inherently qualitative notion of betweenness as primitive, and show how it naturally leads to patterns of interpolative reasoning. Both a semantic and a syntactic characterization of this process is presented, and the computational complexity is analyzed. Finally, some patterns of extrapolative reasoning are sketched, based on the notions of betweenness and parallelism.
KeywordsAttribute Space Rule Base Categorization Rule Conceptual Space Qualitative Reasoning
Unable to display preview. Download preview PDF.
- 6.Gärdenfors, P.: Conceptual Spaces: The Geometry of Thought. MIT Press, Cambridge (2000)Google Scholar
- 7.Gardenfors, P., Williams, M.: Reasoning about categories in conceptual spaces. In: Int. Joint Conf. on Artificial Intelligence, pp. 385–392 (2001)Google Scholar
- 8.Gérard, R., Kaci, S., Prade, H.: Ranking alternatives on the basis of generic constraints and examples: a possibilistic approach. In: International Joint Conference on Artifical Intelligence, pp. 393–398 (2007)Google Scholar
- 11.Lehmann, F., Cohn, A.G.: The EGG/YOLK reliability hierarchy: semantic data integration using sorts with prototypes. In: Int. Conf. on Information and Knowledge Management, pp. 272–279 (1994)Google Scholar
- 13.Prade, H., Schockaert, S.: Completing rule bases in symbolic domains by analogy making. In: 7th Conf. of the European Society for Fuzzy Logic and Technology (2011)Google Scholar
- 16.Schockaert, S., Prade, H.: An inconsistency-tolerant approach to information merging based on proposition relaxation. In: AAAI Conf. on Artificial Intelligence, pp. 363–368 (2010)Google Scholar
- 17.Schockaert, S., Prade, H.: Interpolation and extrapolation in conceptual spaces: A case study in the music domain. In: Proceedings of the 5th International Conference on Web Reasoning and Rule Systems (2011)Google Scholar