Problem Solving

TL;DR: An implementation of the proposed method is implemented and experimental results are reported that support the on-sion that abstra tions are indeed useful for the onsidered task.

Abstract: Several explanation and interpretation tasks, su h as diagnosis, plan re ognition and image interpretation, an be formalized as abdu tive and onsisten y reasoning. The interpretation task is usually exe uted for the purpose of performing a tions, e.g., in diagnosis, repair a tions or therapy. Some proposals address the problem based on a task-independent representation of a domain whi h in ludes an ontology or taxonomy of hypotheses and observations. In this paper we rely on the same type of representation, and we point out the role of abstra tions in an iterative abdu tion pro-ess. Atea h iteration, as in model-based diagnosis and troubleshooting, our algorithm hooses to perform further observations or a tions taking into a ount their osts and the likelihood of andidate hypotheses. The main goal of the algorithm is to ensure dis rimination among hypotheses and, more importantly, to perform the appropriate a tions for the ase at hand. We dis-uss an implementation of the proposed method and report experimental results that support the on lu-sion that abstra tions are indeed useful for the onsidered task.