AAAI Publications, Workshops at the Twenty-Eighth AAAI Conference on Artificial Intelligence

Font Size: 
Extending PSL with Fuzzy Quantifiers
Golnoosh Farnadi, Stephan H. Bach, Marie-Francine Moens, Lise Getoor, Martine De Cock

Last modified: 2014-06-18

Abstract


Probabilistic soft logic (PSL) is a probabilistic modeling framework which uses first-order logic and soft truth values in the interval[0;1] for reasoning in relational domains. PSL uses the Łukasiewicz t-norm and t-conorm from fuzzy logic to model respectively conjunction and disjunction. A PSL rule such as Trusts(A;X)^Trusts(X;B)->Trusts(A;B) models that “A trusts B” is true to the degree to which there is a trusted third party X. In the current version of PSL there is no way to express that A should trust B if most trusted friends of A trust B. In this work, we propose an extension of PSL with fuzzy quantifiers to address this limitation.

Keywords


Probabilistic soft logic (PSL); statistical relational learning (SRL) ;Social network; Fuzzy quantifier

Full Text: PDF