自然语言人机接口中分布智能的研究与设计

传统的自然语言人机接口的主要以串行方式进行处理，各个模块对应于不同的语言层次，这种方法在处理歧义，解决系统的一致性和通用性之间的统一及系统的扩展性上存在较大的局限性。本文提出了一个基于分布式人工智能的自然人机接口模型，利用分布人工智能来实现了一个没有中央控制、通过各个通用语言处理Ａｇｅｎｔ及语用处理Ａｇｅｎｔ交流的自然语言人机接口框架，文中介绍了它的设计它的设计思想和总体结构框架以及Ａｇｅｎｔ之间

Study and Design of Distributed Artificial Intelligence in Natural Language Interface

Natural Language Interface raises challenges for computer to understand sentences that may haveinvalid structure, ambiguity in meaning and area - specific usage. All these problems made conventional NLI sys-tems almost impossidle to become both practical and general - purposed. In contrast to the sequential architectureused in the conventional systems which does not allow real and parallel exchange between different modules, thispaper presents a general - purposed NLI system called SAGE. It adopts Distributed AI in Natural Langunge In-terface. Different modulles are constructed into intelligent agents, which can process the problem with its ownknowledge and interact with others. NLPP is defined as the conversation protocol. This architecture has greatercapability in solving ambiguity than conventional ones, and advantnges like being both practical and general -puapsed,being open for integrating other new methods of natural language processing, and easy to add learningmechanisms to the system.