Abduction and Induction Based on Non-Monotonic Reasoning

摘要: We consider abduction and induction in Artificial Intelligence, focussing on one particular case for each, namely diagnostic learning. Overall, we formalize as well two syntactic specializations of a single reasoning scheme, leading from observed consequences to plausible hypotheses. The problem finding hypotheses that justify given facts is then transformed into an inference, these very facts, prior relevant background knowledge, the corresponding This found be actually form non-monotonic amenable some appropriate logic, where direction reversed: A logic “reversing implication” obtained. give sequent system with application both example induction.