Beginning with introduction of the state of art situation in Cognitive Psychology's Short-term Memory studies, this paper presents the Machine Learning in AI and the Special Noun recognizing method. Furthermore, this paper makes a comparison among them. Though it has been long for Hierarchical Network of Concepts coming into being, there are a few people to know it. This paper gives its outline in chapter 3th. In the following chapter 4~(th) and chapter 5~(th), under the guiding of HNC and with the rational of "from Sentence Category Analysis and for Sentence Category Analysis", the manipulating methods of HNC based Short-term Memory are given. Thinking of the different importance of them for Sentence Category Analysis, the contents, which must been deal with by the Short-term Memory, are divided into three sets as Special Noun, New Chinese Word and the third Set that contains all the other things not been included by the prior two. Based on K-Schedule and Sentence Category Analysis this paper has designed the frame of Short-term Memory and given the flow of control and data. In the final chapter, this paper's summary is stated out. HNC has been being on the progress, so the STM need a lot of works to be done. Because of the limit of time and the author's lack of knowledge there must be many deficiencies in the paper. Any corrections are appreciated.


Key words: natural language process  HNC  Short-term Memory  Special Noun