协同智能计算系统——理论模型及其应用

Collaborative Intelligent Computing System——Theoretical Model with Its Application

本研究工作目的是揭示计算机数据与自然人知识两类信息处理方式基础之上派生的协同智能观及其指导下的协同智能计算系统的理论模型及其应用。它涉及的可验证方法,一方面,在n2矩阵范围内,以等价于2n的发散方式枚举和以等价于1/2n的收敛方式搜索两类基本算法可用作处理纯数字计算的任务,其特征是满足n的取值不影响计算效率的间接计算任务;另一方面,在与n2矩阵各个格子一一对应的范围对单音节汉字进行间接形式化处理,其特征不仅在于单音节字,即言,可间接计算,而且,还在于双音节和多音节的字组,即语,也可间接计算,同时,言和语的复用频率均可以且便于间接计算和统计。其结果是:不仅中文的自然语言理解的双重技术路线被揭示,而且,支配这类间接计算模型与间接形式化方法的信息基本定律假说也可被验证。最终可得出这样的结论,即:在前述两方面可验证的两种实证方法,远不仅仅是计算机数据信息处理方式与自然人知识信息处理方式这两类信息处理方式的简单相加,而是这两者合理分工、高度协作所产生的协同智能计算系统的理论模型或第三类信息处理方式及其应用,例如:国内外学术前沿的各类期刊及会议论文摘要、各种软件的常用问题解答以及帮助文件、协同智能计算系统用户个性化记录、自然语言的有限符号及其多样化组合或重复使用过程中蕴含的有限规则、等等各类双语信息的计算机辅助分析,该类云端计算主要服务对象是在创造性合作型生产式教研产学用各类活动中需要计算机辅助双语知识信息数据处理服务的客户。

This study is aimed at revealing the view of Collaborative Intelligence derived from the mode of two types of Information Processing, namely computer data and human knowledge, as well as the theoretical model of Collaborative Intelligence Computing System with its application guided by this view. It involves a verifiable method. On the one hand, within the range of n2 matrix, the two basic algorithms, which are enumeration based on divergence equivalent to the way of 2^n and search based on convergence equivalent to the way of 1/2^n, can be used as tasks of pure digital computing. It is characterized by satisfying the condition that the value-taking of n does not influence computational efficiency of the Indirect Computational Tasks; On the other hand, within the range of n2 matrix corresponding to each grid, Indirect Formalization Processing is used towards single character, which is characterized not only by that single-Zi-syllable, namely Yan, can be calculated indirectly, but also by that two-Zi-syllable and multi-Zi-syllable, namely Yu, can also be calculated indirectly. At the same time, the frequencies of the reuse of Yan and Yu can both be easily counted and calculated indirectly. Thus, this model can not only lead us to revealing the double technique route toward Chinese Information Processing or Natural Language Understanding, but also to ver- ifying the hypothesis of Informatics Basic Laws, which dominate the Indirect Computing Model and Indirect Formalization Method. Even- tually, it can be concluded that the previous two verifiable empirical methods are much more than a simple sum of these modes of the two types of Information Processing, namely computer data and human knowledge, but the theoretical model of Collaborate Intelligent Computing System or the third kind of Information Processing Mode with its application generated by Rational Division of labor, highly Collaborative Synergy, such as academic forefront of various journals and conference abstracts home and abroad, answers to Frequent Answered Questions of a variety of software and their Help Files, individual records of Collaborative Intelligent Computing System users, limited symbols of natural language and its limited rules inherent in the diverse combination process or the reuse process and so on, which are the Computer-Aided Analysis of all kinds of Bilingual Information. The main target of such cloud-based computing is the customers who need Computer-Aided Bilingual Knowledge and Information Processing in various types of creative cooperative and productive activities, such as teaching, research, production, learning and using.