Search

In this paper, we present a distance-based framework for DL-Lite based on the notion of features. Within this framework, we propose a distance-based paraconsistent semantics for DL-Lite where meaningful conclusions can be rationally drawn even from an inconsistent knowledge base and we develop a distance-based inconsistency measurement for DL-Lite to provide more informative metrics which can tell the differences between axioms causing inconsistency and among inconsistent knowledge. Furthermore, we investigate several important logical properties (e.g., consistency preservation, closure consistency, splitting property etc.) of the entailment relation based on the new semantics and show its advantages in non-monotonic reasoning for DL-Lite. Finally, we show that our two distance-based inconsistency measures are basic inconsistency measures where some good properties hold such as Free Axiom Independence and Dominance of inconsistency etc. [ABSTRACT FROM AUTHOR]

Probabilistic causal interaction models have become quite popular among Bayesian-network engineers as elicitation of all probabilities required often proves the main bottleneck in building a real-world network with domain experts. The best-known interaction models are the noisy-OR model and its generalisations. These models in essence are parameterised conditional probability tables for which just a limited number of parameter probabilities are required. The models assume specific properties of intercausal interaction and cannot be applied uncritically. Given their clear engineering advantages however, they are subject to ill-considered use. This paper demonstrates that such ill-considered use can result in poorly calibrated output probabilities from a Bayesian network. By studying, in an analytical way, the propagation effects of noisy-OR calculated probability values, we identify conditions under which use of the model can be harmful for a network's performance. These conditions demonstrate that use of the noisy-OR model for mere pragmatic reasons is sometimes warranted, even when the model's underlying assumptions are not met in reality. [ABSTRACT FROM AUTHOR]

Abstract: The next development in building Bayesian networks will most likely entail constructing multi-purpose models that can be employed for varying tasks and by different types of user. In this position paper, we argue that the development of a special type of ontology to organize the knowledge involved in such a multi-purpose model is crucial for the management of the model’s content. This ontology should preserve all knowledge elicited for the construction of the model and be accessible to domain experts and knowledge engineers alike. Based on the different ways in which people learn and gain expertise, we further argue that knowledge elicitation will result in task-specific knowledge mostly, which is best stored in the format in which it is elicited. To support varying model views for different tasks and different types of user, we propose that the elicited knowledge be organized in a library-style ontology of separate modules. [Copyright &y& Elsevier]