Relaxed Abduction

Thomas Hubauer addresses the challenge of providing reasonable interpretations of incomplete observational data in the context of imperfect domain models - a situation often encountered in the context of industrial diagnostics. To tackle this problem, the author proposes a novel approach called Relaxed Abduction, which is able to derive pragmatic interpretations in situations where existing methods either fail or provide overly complex solutions. To strengthen the link to applications in industrial diagnostics, he develops a methodology to structure diagnosis problems according to ISO 13379 and express it using multiple description logic knowledge bases.ContentsFormalizing Diagnostics Using Description LogicsFormalization and Properties of Relaxed AbductionSolving Diagnostic Problems Using Relaxed AbductionIncremental Diagnostics Target GroupsResearchers and students of computer science, knowledge management, and applied mathematicsPractitioners in the area of industrial diagnostics and maintenanceThe AuthorThomas Hubauer is a Research Scientist at the R&D department of a large German engineering company. His topics of interest range from ontologies and reasoning to machine learning and data analytics.