LOGICAL REASONING VIA SATISFIABILITY MAPPED INTO ENERGY FUNCTIONS
Priscila M. V. Lima, Mariela Morveli-Espinoza, Glaucia C. Pereira, T.O. Ferreira, Felipe M. G. França · International Journal of Pattern Recognition and Artificial Intelligence · 2008
This paper presents the implementation of ARQ-PROP II, a limited-depth propositional neural reasoner based on the Resolution Principle. The SATyrus platform was used in the synthesis of Energy functions from a set of pseudo-Boolean constraints specifying ARQ-PROP II architectures for different inferencing depths. Global minima of the Energy functions produced by SATyrus are associated to SATisfiability of a formula and, in the case of ARQ-PROP II, are associated to Resolution-based refutations. This allows for simplified abduction, prediction and planning to be unified with deduction in a goal-driven style, i.e. there is no need for presetting a reasoning style upon a target set of clauses. Experimental results on deduction with ARQ-PROP II using different propositional depth settings are presented together with a correction of Gadi Pinkas' mapping of SATisfiability into Energy minima.