TABLING WITH INTERNED TERMS ON CONTEXTUAL ABDUCTION

Abstract

Abduction (also called abductive reasoning) is a form of logical inference which starts with an observation and is followed by finding the best explanations. In this paper, we improve the tabling in contextual abduction technique with an advanced tabling feature of XSB Prolog, namely tabling with interned terms. This feature enables us to store the abductive solutions as interned ground terms in a global area only once so that the use of table space to store abductive solutions becomes more efficient. We implemented this improvement to a prototype, called as TABDUAL+INT. Although the experiment result shows that tabling with interned terms is relatively slower than tabling without interned terms when used to return first solutions from a subgoal, tabling with interned terms is relatively faster than tabling without interned terms when used to returns all solutions from a subgoal. Furthermore, tabling with interned terms is more efficient in table space used when performing abduction both in artificial and real world case, compared to tabling without interned terms.