The Theory of Bio-Medical Knowledge Integration(Ⅵ)

 

                 Hanfei Bao 
 
  Abstract   This paper introduced the following new concepts：the cognitive goal, the cognitive goal for the declarative data of the patient records (PRs), The basic attributes of PR’s data at the sides of generation, construction and cognition, the generalized data creator (GDC), type Ⅰ to Ⅵ+ of GDC, the cognitive directions of data: forward direction and backward direction, the apparent cognitive orientation and inapparent cognitive orientation, the cognitive granularity difference principle between the natural intelligence and the artificial intelligence, the generalized variable(GVAR) and the generalized value(GVAL), the variable and value transitivity law(V-V transitivity law),  the attribute-combination irreversibility between the concept abstracting and embodying, an open model of the launching engine of bio-medical cognition,  etc

： 

.

.

 

 

摘要  本文介绍如下新概念:认知目标,病历描述性数据的认知目标,病历数据的基本属性:生成属性,表达属性,认知属性,广义数据生成器(GDC):Ⅰ到 Ⅵ+型GDC,数据的认知方向:顺向认知方向，逆向认知方向,显性认知方向，隐性认知方向，自然智能与人工智能的认知粒度差异原理，广义变量，广义变量值及广义变量-广义变量值（在本体结构中的）传递律（V-V transitivity law），抽象过程与指定过程定义属性组合自由性的不可逆性，生物医学认知行为激发器模型等。

 

 

 

此帖由 老包 在 2005-12-09 11:17 进行编辑...

  

1．A work independent of the particular works of knowledge engineering in BioMedicine

 

The Theory of BioMedical Knowledge Integration(BMKI)^[1-16] considers neither much upon the aspects of syntactic level of information (such as rule-, frame-, object-, ontology-, text-, image-, progress- or video-based knowledge forms), which being the hottest spots nowadays in the topic of knowledge support in the Medical Informatics, nor upon the development of knowledge or data bases themselves. BMKI focuses very much herself on the semantic behaviors of data, information and knowledge. Supposing one day all the phases of the issues on the information storage, communication, showing, etc  ie  all those of information-syntax-level, had been fully or perfectly resolved, or in other words, the diverse heterogeneous or hetero-structured data and knowledge bases reciprocally and ideally linked or communicated, or, if I can say, that all the original information always presented to the users according to their wishes and ready to be used, what kind of challenges on the medical knowledge processes would have occurred to them.  What sorts of the black holes would have waited for them. Those considerations make the sights of the Theory of BioMedical Knowledge Integration(BMKI) having constantly or invariably been concentrated on such a mysterious scene. Therefore, it may be said, in a certain sense, BMKI is independent of many activities of the knowledge or data operations in medical informatics today, but relevant to their future. 

 

 

               1. 独立于生物医学知识工程具体开发的研究

BMKI考虑的重点既不是作为当前医学信息学领域中知识支持研究的热点的信息语法层次的（如规则的、框家的、对象的、本体论的、文本的、图象或图形的、过程的或电视的）知识表达形式，也不是知识库和应用程序的开发和研究本身。BMKI的关注点为数据、信息和知识的语义行为。假设有朝一日关于信息的存储、通讯、显示等所有的问题也即所有的信息语法层次的问题都已得到充分的解决，换言之，各种异源异构的数据库和知识库都能畅通无阻地交互连接和传输，或者说所有的原始信息能按照用户的希望提呈给用户，那么用户将面临何种新的挑战呢？用户将面临何种知识黑洞呢？BMKI始终如一地将她的目光投射到这个神秘的区域。所以在某种程度上，BMKI独立于当前的一些医学信息学的知识或数据领域的探索，而与其未来有关。

摘自:HF Bao:The Theory of Bio-Medical Knowledge Integration(Ⅵ)

说明: 概念抽象化（Abstraction）和具体化（Embodying）是知识本体或其他任何知识体最根本的最常见的智能行为，我们不妨称之为A-E运动（A-E motion）。一个较为清晰或明确的知识单元按惯例一般都会解析为主体（或资源）-特征（属性或变量或谓词）-特征值（或变量值或客体）三元体。BMKI更愿意把它表为逻辑上或数学上或工程上更为基础的更可操作的“主体-变量-变量值” 三元体。BMKI根据自己独立的长期思考在A-E运动中提炼出所谓“变量-变量值角色传递律”（Variable and value transitivity law，V-V 传递律)）。后来发现OWL等语言在这个问题上也有个别经验性的说明，但并未意识到其理论意义。现把V-V 传递律表述如下：.

.

 

4.The abstracting and embodying of concept.

Abstraction of concept is a process to get meta-concepts or meta-class, through continuing to take some attribute or their combination away, freely, from the definitional attribute set(DAS) of the original concept or class(see Fig. 3). Contrarily, Embodying  of concept means to specify the sub-concepts or the instances, by means of adding some attribute or their combination, not so freely, to the DAS of the original concept or class(see Fig. 4).
In the courses of abstraction and embodying, there are two important concepts involved:
Generalized Variable(GVAR): The aspects of observation or description of thing.
Generalized Value(GVAL): The specification of the GVAR.
In these courses, things change their roles between GVAR and GVAL. For examples, “(GVAR)which type of student? (GVAL)medical student” → “(GVAR) which type of medical student? (GVAL) medical student majoring pediatrics” →……(see Table 3).
Then we come to:
   Variable and value transitivity law(V-V transitivity law): Let Y_i+1 = f(Y_i)）is a hierarchical tree by means of is_a relationships, Y_i+1 is a subclass or instance of Y_i, i = 1,2,3,…. then when Y_i is applied as a generalized variable(GVAR), Y_i+1 would be its generalized value(GVAL). And the same for Y_i+1 and Y_i+2……

   More formalized, let A or B be a concept or GVAR or GVAL and “A → B” means (A is a (meta-)GVAR of B)∨(B is a (sub-)GVAL of A),then (meta-)ⁿconcept is a (meta-)GVAR of (meta-)^n-1concept)∨((meta-)^n-1concept is a (sub-)GVAL of (meta-)ⁿconcept)..

.

 

 

4.概念抽象化与具体化.

概念的抽象化是获得原概念的元概念或元类的过程，通过不断地自由地从原概念的定义属性集合（The Definitional Attribute Set，DAS）中抽去某种或某些属性以实现此过程（见图3）。相反，概念的具体化为指定原概念的子概念或实例的过程，通过不断地但并非完全自由地或不无约束地向DAS中增加某种或某些属性以实现此过程（见图4）。
在抽象化和具体化过程或A-E运动中，涉及二个重要概念：
广义变量（Generalized Variable，GVAR），即对事物观测或描述的方面；
广义变量值（Generalized Value，GVAL)，即对GVAR的指定。
   在A-E运动中，事物往往在GVAR和GVAL之间转变角色。例如，“问：什么类型（GVAR）的学生？答：医学生（GVAL）” →“问：什么类型的医学生（GVAR）？答：儿科学学生（GVAL）” →……（见表3）。
把问题一般化，我们就得到：
变量-变量值角色传递律：
设Y_i+1 = f(Y_i)）为由is_a 关系形成的层次树结构，Y_i+1为Y_i, i = 1,2,3,…的子类或实例，于是当Y_i被用作GVAR，Y_i+1可被用作GVAL，对Y_i+1 and Y_i+2……则依次类推。
更形式化地，设A或B为一个概念或GVAR或GVAL，“A → B”意为 “A为a (meta-)GVAR of B)∨(B为a (sub-)GVAL of A)”,那么“(meta-)ⁿ概念→(meta-)^n-1概念”意为“(meta-)ⁿ概念为a (meta-)GVAR of (meta-)^n-1概念)∨((meta-)^n-1概念为a (sub-)GVAL of (meta-)ⁿ概念”。.

.

 

摘自Hanfei Bao:The Theory of Bio-Medical Knowledge Integration(Ⅵ)(accepted by <Medical Information>)

 

 

此帖由 老包 在 2006-01-08 14:37 进行编辑...