一个面向图形/图像的机器思维与理解模型

A Graphics/Image Oriented Model for Artificial Intellegent and Understanding

本文通过对句法结构模式识别法以及人脑形象化思维方式的研究,提出一种面向图形/图像的知识表示与思维模型。这种模型以图形/图像的方式来表示知识,通过对输入图形/图像的反复交换、识别与理解,得到问题的描述。本文给出了该模型的数学定义,研究了其中的变换运算,并用一个场景对话实例作出了初步实践。

A new graphics/image oriented model for knowledge representation and reasoning is proposed. Based on syntax pattern recognition and human reasoning model, knowledge is first described as graphs and images. After reiteratively tansforming, recognizing and understanding of those input graphs and images, a desciption of the problem is obtained. The mathematics definition of this model is given. Its related transforms are proposed and analysed. Experimental results are provided in a 'scenerio dialogue' test.