4 February 2011 , Computational Social Choice Seminar, Daniele Porello
The problem of aggregating different individual ontologies into a collectively accepted one has been discussed in particular in connection with belief merging. In this talk, I will present some results on the application of methods developed for voting theory and judgment aggregation to the problem of aggregating formal ontologies.
We will discuss mainly two different scenarios: (1) individuals agree on the definitions of concepts but may disagree on factual assertions, and (2) individuals disagree also on the definitions of concepts. We will see that the plurality rule does not always provide a consistent ontology, since instances of the discursive dilemma can easily be found for description logic languages. Then we will see on which ontologies the plurality rule provides consistent outcomes and we will discuss other voting procedures (mainly distance-based procedures) for the cases in which plurality fails.
This is joint work with Ulle Endriss.