Imperfect Information Flow: Modelling Channels and Defaultsin Domain Theory

William C. Rounds University of Michigan $84,231 - 12 most。不完全信息流：领域理论中的建模通道和默认值这项工作是尝试使用编程语言语义中使用的技术来统一智能系统建模中的一些概念。更具体地说，它研究的是允许错误但仍然允许有用信息传递的信息流。在以前的工作中，已经发展了一种基于领域理论的这种信息流理论，称为默认模型理论。默认模型理论允许在新的“严格”信息可用时重写计算中某一点使用的信息，从信息通道的角度提供了一种理解默认和非单调推理的方法。这项工作还将研究在自然语言处理、逻辑编程和数据库处理方面的应用。

IRI-9509067 William C. Rounds University of Michigan $84,231 - 12 mos. Imperfect Information Flow: Modeling Channels and Defaults in Domain Theory This work is an attempt to use techniques used in the semantics of programming languages to unify some concepts in intelligent systems modeling. More specifically, it studies the flow of information which admits errors but still allows useful information to be transmitted. In previous work, a theory of such information flow, called default model theory , grounded in domain theory, has been developed. Default model theory allows information used at one point in a computation to be overwritten when new "strict" information becomes available, providing a means of understanding default and nonmonotonic reasoning from the perspective of information channels. This work will also investigate applications to natural language processing, logic programming, and database handling.