Integrating L m4c with Abductive Reasoning for Chance Specification and Chance Discovery Inference

Since the term was coined by Ohsawa, Chance Discovery has become an active research area. The Chance Discovery research is involved in the following two basic questions: how to define a chance and how to discover chances. Abe actually have answered these two questions by introducing Abductive Reasoning (AR) and then extending to Abductive Analogical Reasoning (AAR). His explanation is very close to the intuitive meaning of Chance Discovery according to some philosophical tradition. However, Abductive Reasoning still has to be extended in the following two dimensions: (1) From selective perspective to creative perspective; (2) From explanatory perspective to planning perspective. In this paper, we will propose a solution to above two basic questions by integrating Abductive Reasoning with Lm4c, a formal theory established to characterize the influence of motivational attitudes. As a result, Abductive Reasoning is extended in the above dimensions. Due to some excellent properties of L m4c, we achieve some desirable effects by the integration. First, the hypotheses can be generated automatically from the knowledge base Σ and the desired goal G. This generation requires no pre-given hypotheses any longer, and all the hypotheses automatically generated can be explained reasonably in some intuitive semantics. Second, the amount of the elemental chances automatically generated is finite, among which only the core ones are left. Provided with this set of chances, an agent can concentrate his/her attention on it. Based on these two points, we can conclude that the integration of Abductive Reasoning with Lm4c provides a more powerful and elegant theory for both the specification of Chance and the inference in Chance Discovery.