Increasing complexity of technical systems requires a precise fault localization in order to reduce maintenance costs and system downtimes. Model-based diagnosis has been presented as a method to derive root causes for observed symptoms, utilizing a description of the system to be diagnosed. Practical applications of model-based diagnosis, however, are often prevented by the initial modeling task and computational complexity associated with diagnosis. In the proposed thesis, we investigate techniques addressing these issues. In particular, we utilize a mapping function which converts fault information available in practice into propositional horn logic sentences to be used in abductive model-based diagnosis. Further, we plan on devising algorithms which allow an efficient computation of explanations given the obtained models.
This paper presents a novel approach to diagnosis which addresses the two problems - computational complexity of abduction and device models - that have prevented model-based diagnostic techniques from being widely used. The Experience-Aided Diagnosis (EAD) model is defined that combines deduction to rule out hypotheses, abduction to generate hypotheses and induction to recall past experiences and account for potential errors in the device models. A detailed analysis of the relationship between case-based reasoning and induction is also provided. The EAD model yields a practical method for solving hard diagnostic problems and provides a theoretical basis for overcoming the problem of partially incorrect device models.
Most work in symbolic concept acquisition assumes a deductive model of classification in which an example is a member of a concept if it satisfies a logical specification represented in disjunctive normal form (DNF) (Michalski and Chilausky, 1980)) a decision tree (Quinlan, 1986), or a set of Horn clauses (Quinlan, 1990). However, recent research in diagnosis, plan recognition, object recognition, and other areas of AI has found that abduction, finding a set of assumptions that imply or explain a set of observations, is frequently a more appropriate and useful mode of reasoning (Charniak and McDermott, 1985; Levesque, 1989). This paper concerns inducing from examples a knowledge base that is suitable for abductive reasoning. We focus on abductive diagnosis using the model of (Peng and Reggia, 1990).
Abduction is a basic form of logical inference, which is said to engender the use of plans, perceptual models, intuitions, and analogical reasoning - all aspects of Intelligent behavior that have so far failed to find representation in existing formal deductive systems. This paper explores the abductive reasoning process and develops a model for it s mechanization, .which consists of an embedding of deductive logic in an iterative hypothesis and test procedure. An application of the method to the problem of medical diagnosis is discussed.In IJCAI-73: THIRD INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, 20-23 August 1973, Stanford University Stanford, California.
The paper introduces a logical framework for negotiation among dishonest agents. The framework relies on the use of abductive logic programming as a knowledge representation language for agents to deal with incomplete information and preferences. The paper shows how intentionally false or inaccurate information of agents could be encoded in the agents' knowledge bases. Such disinformation can be effectively used in the process of negotiation to have desired outcomes by agents.