Proceedings Abstracts of the Twenty-Third International Joint Conference on Artificial Intelligence

Syntactic Computation of Hybrid Possibilistic Conditioning under Uncertain Inputs / 739
Salem Benferhat, CĂ©lia da Costa Pereira, Andrea G. B. Tettamanzi

We extend hybrid possibilistic conditioning to deal with inputs consisting of a set of triplets composed of propositional formulas, the level at which the formulas should be accepted, and the way in which their models should be revised. We characterize such conditioning using elementary operations on possibility distributions. We then solve a difficult issue that concerns the syntactic computation of the revision of possibilistic knowledge bases, made of weighted formulas, using hybrid conditioning. An important result is that there is no extra computational cost in using hybrid possibilistic conditioning and in particular the size of the revised possibilistic base is polynomial with respect to the size of the initial base and the input.