Belief base contraction by belief accrual

Merging existential rules programs in multi-agent contexts through credibility accrual

“…To do this, we will link formulas to certain values calculated from the credibilities associated in the community to the agents that are the informants of such formulas. This calculation will be made by exploiting particular accrual functions [13] tailored to our multiagent scenario.…”

“…Nevertheless, Kernel Contraction does provide the foundations for how our approach deals with conflicts, since we define the merging process as the application of incision functions on the union of Datalog ± ontologies while still considering the different sources of the formulas by means of accrual functions. In particular, instead of directly considering minimal inconsistent subsets of formulas in the ontology (which are equivalent to ⊥-kernels), in this work we consider incisions over structures called clusters [37,12,13] that groups together related kernels.…”

“…Once we have found the conflicts, the next logical step is to solve them somehow. To do so, as explained in Section 2 in this work, we will exploit the findings in [13], defining accrual functions as a way to measure the credibility associated with a particular formula, based on the credibilities attached to those agents that claim said formula.…”

“…While this design decision has several advantages, like greater consensus and bias removal, the decentralization of knowledge maintenance could also have disadvantages, in particular regarding conflicting pieces of information leading to inconsistencies in the knowledge. Through the years many proposals dealing with such issues were introduced, ranging from the field of AI [6,7,11] and Belief Revision [1,13,24] to that of Databases [5,33,38], to cite a few of an extensive literature.…”

“…Our proposal is based on credibility accrual functions; these functions will assign values to formulas based on the agents' credibility value that inform them. Then, to resolve information conflicts, we define incision functions based on accrual functions [24,13] that select which formulas should be removed from the conflicting sets. The ontology resulting from this process is the set obtained by integrating all formulas that are not removed by these incision functions.…”