摘要： 将神经网络与逻辑推理统一到面向对象理论中，建立了同时具备神经网络和专家系统特性、融合连接机制和符号机制的智能神经元模型。提出了一种新的程序设计语言——智能神经网络程序语言(NIPL)，实现了神经计算、逻辑推理和数值计算的统一。定义了NIPL的语法，设计并实现了智能神经网络程序设计语言NIPL编译器，从而为开发智能神经网络应用系统提供了有效的手段。

关键词: 智能神经元模型, 智能神经网络程序设计语言, 抽象语法, NIPL编译器

Abstract: As ideas of neural networks and logic inference are introduced into the theory of object-oriented, an intelligent neuron model is built. This model, with the characteristic of neural networks and expert system, integrates the link mechanism with the symbol mechanism. Then a new programming language——intelligent neural network programming language called NIPL is presented. NIPL makes the unification of neural network computation, logic inference and numerical computation. The abstract grammar of NIPL is given. The compiler of NIPL is designed and implemented and thus an effective means for development of the intelligent system is provided.

Key words: Intelligent neuron model, Intelligent neural network programming language (NIPL), Abstract grammar, NIPL compiler