Reference: McIlraith, S. A. Logic-Based Abductive Inference. Knowledge Systems Laboratory, July, 1998.
Abstract: This paper surveys the work on abductive inference within the field of artificial intelligence (AI), with particular attention to logic-based abduction. The paper commences with a formal description of three popular characterizations of abductive inference. This is followed by an examination of several specific logic-based abductive frameworks, each of which applies syntactic restrictions to the formulation of the abductive reasoning problem and the resultant explanation. Mechanisms for computing logic-based abductive explanations, and the complexity of variants of the abduction task are examined in the sections to follow. This paper also surveys different applications of abduction in AI, and the connections between abduction and other types of nonmonotonic reasoning. The paper concludes with a discussion of potential future research areas.
Full paper available as ps, ps.