關(guān)于 ontology 的討論

1、關(guān)于 ontology 的討論董云衛(wèi)： 你問道關(guān)于ontology, 直譯是哲學(xué)上的存在論或本體論，現(xiàn)在用于系統(tǒng)的概念模型很熱。大體意思是說，客觀世界是由很多元素組成，而元素之間又具有各種聯(lián)系，把這些元素和關(guān)系畫出來就是一個(gè)ontology。這里有幾篇文章可供參考，都是2002年國(guó)際會(huì)議的文章，比較新。 1， 25030001 Conceptual Modeling and Ontology: Possibilities and Pitfalls 2， 25030003 An Ontology for m-Business Models 3, 25030012 Ontology-Driven

2、Conceptual Modeling: Advanced Concepts 4, 25120174 DAML+OIL:A Reason-Able Web Ontology Language 郝克剛。2003年1月15日Sent: Wednesday, January 15, 2003 8:49 PM Subject: RE: 關(guān)于ontology 郝老師及各位： The 10th international human-computer interaction conference (2003) 剛接受了我一篇文章：Ontology-based Conceptual Modeling for

3、 Interaction Design. 創(chuàng)新和質(zhì)量都得了滿分:-) 事實(shí)上，其內(nèi)容是討論軟件系統(tǒng)的概念 建模的，但和建模交互有一定關(guān)系。我和董云衛(wèi)爭(zhēng)論過2小時(shí)，但他不相 信我的。 本體論的常用定義是：分享概念化的形式、顯式規(guī)約（但有爭(zhēng)論），其 內(nèi)容包括一個(gè)概念分類，關(guān)系及公理。本體論一般是靜態(tài)的，不包括動(dòng)態(tài) 概念。換言之，本體論描述的是說明式知識(shí)，不包括過程序知識(shí)，因?yàn)?本體論的目的是表示，不是使用知識(shí)。所謂分享概念化指是在一個(gè)問題域 中現(xiàn)象的抽象模型，其中概念是公認(rèn)的，形式化指機(jī)器可處理性， 顯式指概念的類型和使用限制都是明確定義的（一般得有一個(gè)meta-ontology 或叫ontolo

4、gy assumptions定義概念類型和類型之間的關(guān)系，一個(gè)具體的 概念模型中的概念及關(guān)系是它的實(shí)例）。 祝新年快樂！ 華慶一Sent: Thursday, January 16, 2003 5:21 AMSubject: 關(guān)于ontology 郝老師及各位, 我最近在看一些Semantic Web方面的資料, 郝老師提到的DAML+OIL語(yǔ)言屬于 Semantic Web 語(yǔ)言, 用于描述Web document and application類及其之間的關(guān)系. Ontology 本是哲學(xué)用語(yǔ), 現(xiàn)在廣泛用于其它領(lǐng)域的關(guān)系建模. 下面是SemanticWeb中 Ontology的定義. 這

5、是由Web 的締造者, 也是Semantic Web的發(fā)起者TimBerners-Lee 在文章”The Semantic Web”中給出的定義. 我的理解是: Ontology 是特定領(lǐng)域的一個(gè)信息集合, 建模該領(lǐng)域?qū)嶓w類及其關(guān)系,以及定義類之間的一組推理規(guī)則. Semanic web中用Ontology 來描述Web Page, 目的是便于Software Agent 自動(dòng)智能查找特定信息, 以及調(diào)用Web application. "ontology is a theory about the nature of existence, of what types of thin

6、gs exist; ontology as a discipline studies such theories. Artificial-intelligence and Web researchers have co-opted the term for their own jargon, and for them an ontology is a document or file that formally defines the relations among terms. The most typical kind of ontology for the Web has a taxon

7、omy and a set of inference rules. The taxonomy defines classes of objects and relations among them. " Best Regards, Yang Hongli發(fā)信時(shí)間: 2003/01/16 09:03主題: RE: 關(guān)于ontology Sorry, I cannot all agree Hongli' understanding. Tim Lee is only a technical guy, and therefore, his description is made fo

8、r his idea of semantic web. DAML and OIL stems from AI, and they are ontology languages for the representation of domain knowledge. of course, they can be used for the derscription of so-called meta- language implicit in the Web and that is the purpose of DAML+OIL which inteperetes the description i

9、nto RDF description. the problem for Lim Lee is who can make a universal ontology for any things from any views for a site called :-) and any machines can find terms from there and then interoperate each other. From my point of view, this is a dream. CYC projects has spent effort for this

10、 more than 10 years! do you had used it once? Problem-solving capability belongs to us, not agents even intelligent ones. we process all ill-defined problems, but agents can process only well-defined ones. And we have to distinguish the things desribed with an ontology from the terms used to define

11、the ontology. As I mentioned earlier, the terms are types or classes, and their relationships about the concepts of someproblem domain of interest. For the purpose of semantic web, ontologies can beused to specify meta knowledge for web pages, such therms as author, location, etc. this is only one u

12、sage of ontology, that is, for semantic interoperation between machine agents. However, who will be respensible for specifying the attributes of those concepts, or called annotation, web-master?:-) and will youagree the concepts within the annotation with the web-master? ontology has a number of the

13、 other usages. for conceptual modeling, the terms are used to capture the user's problems in some domain of interest.Qingyi Hua 華、楊、董等： 看到你們關(guān)于 ontology 的討論，我感到非常高興。希望繼續(xù)深入討論下去。華慶一，祝賀你的文章被高分錄用，有無電子版寄來看看，共大家分享。楊紅麗，你手頭有無關(guān)于 ontology 和 Semantic Web方面電子版的資料或網(wǎng)址供大家學(xué)習(xí)參考。 郝克剛 Hao,kegang 2003年1月17日發(fā)信人: yang

14、 .sg發(fā)信時(shí)間: 2003/01/18 15:09主題: On Semantic WebHi, Everyone, I am glad to have a chance to introduce Semantic Web to you all./ is the wed site of W3C, which responsible for HTML andXML technologies. You can find standards of Semantic Web from there.Attached file is

15、 just the general introduction on Semantic Web.As Hua Laoshi said, Semantic Web seems a dream. It still many of issues onit, such as " harder for users to check to see whether the ontology someoneelse designed has the meaning they expect". There are a number of researchareas that are opene

16、d up by ontology languages, such as DAML+OIL and DAML-S,including issues of tool design, language display, and the use of theontologies by more complex web agents. Discussion of these issues andexamples of the way DAML+OIL is addressing them can be found on the.Actually, I am new b

17、eginner of ontology, now my work is on the semantics ofXML family languages such as XSLT, XPath, XQuery, and XML Web Servicelanguages like WSDL, SOAP and UDDI. The purpose is to specify theinterdependence between these languages. Semantic Web is different storybecause of ontology, but it builds on t

18、he XML technologies. Particularly,Semantic Web service which is described by DAML-S, is high level descriptionof Web service at present. It is the reason why i am interested on SemanticWeb. The questions of Hua Laoshi mentioned are really valuable. Hopefully todiscuss these issue late when I fully t

19、ouch corresponding field.Best regards,Yang Hongli發(fā)信人: Qingyi Huahuaipsi.fhg.de發(fā)信時(shí)間: 2003/01/17 11:02主題: RE: 回復(fù): 關(guān)于ontology郝老師：我的文章正在做最后的修改，等完成后即寄，主要是改的好懂些，這里的人看完后說太生澀難懂，因?yàn)橛昧舜罅啃睦韺W(xué)，知識(shí)工程，和人機(jī)交互的概念。但事實(shí)上，OO中大量概念也來源于心理學(xué)，如繼承，封裝，聚合等。主要目的是如何減少設(shè)計(jì)者的認(rèn)知負(fù)擔(dān)，有時(shí)這種方法稱為system-oriented。用戶為中心的設(shè)計(jì)是user-oriented,目的是減少用戶的認(rèn)知負(fù)

20、擔(dān)，同時(shí)設(shè)計(jì)方法不能改變（設(shè)計(jì)者的負(fù)擔(dān)不能增加。）這是我和董云衛(wèi)的分歧所在。關(guān)于本體論，Ontology (with a capital O)和ontology (with a lowercase o)是有區(qū)別的。前者是指哲學(xué)本體論，難以定義，因?yàn)橛^點(diǎn)不同。后者是人工智能的本體論，可以有復(fù)數(shù)。在哲學(xué)上，Ontology is the study of the KIND of things that exist.在AI中，ontology 常常有兩種用法：one is that ontology is a representationvocabulary that captures concep

21、tualization in some domain of interest.Another is that ontology is used to refer to a body of knowledge, such asCommonsense knowledge by use of a vocabulary, such CYC project.Qingyi Hua發(fā)信人: Qingyi Huahuaipsi.fhg.de發(fā)信日期: 2003/01/21 13:29主題: RE: 關(guān)于ontologySee /maybe much more directly and purposefully:-)