An AI criterion for an account of inference: How to realize a task

We point to two different motivations underlying investigation of relations between various types of inference. The former considers inference as an evidential process and the latter as a methodological process. Furthermore, the approaches in the former group classify inferences differently as they utilize different criteria such as ‘validity’ or ‘value of productiveness’ of conclusions. In AI terminology, the evidential approach aims at constructing an ontology of inferences on the basis of characteristics of the conclusions they draw, without regard to the type of process they involve, while the methodological approach is concerned with the characteristics of these processes. From the AI perspective and particulary in the context of knowledge based systems, the answer to the question of how closely related abductive and inductive inferences depends on the issue of to what extent they do utilize the same set of primitive methods in order to be realized. We model ‘from effects to cause’ type of inference on the basis of case-based reasoning paradigm and utilize contextual information as a means for focusing the generation of hypotheses.