论文部分内容阅读
本文介绍一种建立知识库的学习方法。从人类专家吸取知识的知识获取系统有两种类型:交互式和非交互式的。本文介绍非交互式知识获取系统,它依据观察从人类专家获得知识。这种知识获取系统学习人类专家用以解决问题的策略和从瞬时序列数据中抽取逻辑规则,其学习方法是基于翻译的学习(IBL),在学习过程中要用到预备知识。IBL 有两个子系统:翻译系统和学习系统。前者负责把现实世界的信息翻译成内部规则形式,而规则维护系统负责生成和界定知识。本文将介绍翻译系统和规则生成的预处理过程。
This article describes a way to build a knowledge base. There are two types of knowledge acquisition systems that draw knowledge from human experts: interactive and non-interactive. This article describes a non-interactive knowledge acquisition system that draws knowledge from human experts based on observations. This knowledge acquisition system learns the strategies human experts use to solve problems and to extract logical rules from instantaneous sequence data. The learning method is based on translational learning (IBL), which uses the preparation knowledge. IBL has two subsystems: translation system and learning system. The former is responsible for translating the information in the real world into an internal form of rules, while the rules maintenance system is responsible for generating and defining knowledge. This article describes the preprocessing of translation systems and rules generation.