原站英文机器转录 + 本地中文机器初译(未完成逐句听审) 原作者及许可见 attribution.txt;中文由九天新增,不代表原作者认可。 00:00:10.740 EN: Then welcome for the lecture knowledge engineering in semantic web. Today we talk about OWLs. So you see here an OWL, OWLs because they are wise and they convey wisdom. 中文初译: 然后欢迎在语义网中进行讲学知识工程。今天,我们谈论OWL。你在这里看到一个“OWL ” , OWL , 因为他们是聪明的,能传达智慧。 00:00:25.660 EN: So we talked in the last lectures about RDF, RDF schema, and OWL builds on top of RDF schema. If you remember, RDF schema was basically restricting the use of or providing some structure for triples, for statements. 中文初译: 于是我们在上次的讲座上谈到“RDF ” 、 “RDF ” 和“OWL ” 在“RDF ” 制式之上。如果你记得,“RDF ” 制式基本上限制使用或提供某种结构,供三重发言使用。 00:00:49.000 EN: And OWL is further adding more structure and more restrictions, more logical axioms to knowledge bases, knowledge graphs in RDF. 中文初译: 而“OWL”又增加了更多的结构和更多的限制, 更符合逻辑的知识基础, 知识图中的“RDF”。 00:01:06.670 EN: OWL stands for web ontology language. And you see there the term ontology. Ontology is actually a very old term. 中文初译: “OWL”代表网络本体论语言。你可以看到“OWL”一词。“OWL”实际上是一个非常古老的术语。 00:01:15.910 EN: So it comes from philosophy and was traditionally only used in a singular, in singular. So in philosophy, there are no ontologies because there was only one model of the world, basically, or the idea was that ontology is like a model of the world. 中文初译: 因此,它来自哲学,传统上只用于单一的、单一的。因此在哲学中,没有本体论,因为世界上基本上只有一个模式,或者说,本体论就像世界的模型。 00:01:37.430 EN: Now in computer science, we actually have many ontologies. We can build lots of ontologies. Ontology in philosophy meant the science of being. 中文初译: 现在,在计算机科学中,我们实际上有许多本体论。我们可以建立许多本体论。哲学中的本体论意味着存在的科学。 00:01:50.610 EN: And you can find it already in Aristotle, Socrates, Thomas Aquinas, Descartes, Kant, Hegel, Wittgenstein, Heidegger, Quine. 中文初译: 你可以在亚里士多德、苏格拉底、托马斯·阿奎纳斯、笛卡尔、坎特、赫格尔、维特根斯坦、海德格、奎恩等地找到它。 00:02:01.670 EN: So lots of, almost all philosophers have also said something about ontology, the science of being. In computer science, the term was coined in 1993 by Tom Gruber. 中文初译: 所以很多,几乎所有的哲学家 也都对本体论说了一些东西。在计算机科学中,这个词是1993年汤姆·格鲁伯创建的。 00:02:18.480 EN: And his definition of ontology was that ontology is a formal specification of a shared conceptualization of a domain of interest. 中文初译: 他对本体论的定义是 本体论是对利益领域共同概念化的正式说明。 00:02:29.700 EN: And it means that an ontology must be machine interpretable, that it's based on some consensus. So it cannot just be the viewpoint of one particular person, but it should be a consensus of a community of people being experts in a certain domain. 中文初译: 这意味着本体论必须是机器可解释的, 它基于某种共识。所以它不能只是某个特定的人的观点, 但它应该是一个群体在某一领域专家的共识。 00:02:49.340 EN: The ontology describes concepts which are related to a topic, a subject matter. So this is the traditional definition of ontology in computer science. 中文初译: 本体论描述了与一个专题,一个主题事项有关的概念。这就是计算机科学中本体论的传统定义。 00:03:01.400 EN: And since 1993, ontologies gained more and more traction, become more and more popular for organizing knowledge, data, information in various domains. 中文初译: 自1993年以来,本体论获得越来越多的牵引力, 越来越受人欢迎, 在各个领域组织知识、数据、信息。 00:03:17.720 EN: So it's a practical definition. There are some requirements. So there is, for example, the instantiation of classes by individuals. 中文初译: 所以这是一个实际的定义。有些要求。例如, 个人对分类的即时反应。 00:03:29.220 EN: We have already seen that, how that works with RDF schema. We can create hierarchies, taxonomies with inheritance of classes or terms. 中文初译: 我们已经看到了,RDF的策略是如何运作的。我们可以建立等级、分类以及阶级或条件的继承。 00:03:42.760 EN: Also, this we have already seen in RDF schema, where we have classes and subclasses. There are binary relations between individuals. 中文初译: 此外,我们已经在“RDF”计划中看到了这一点,我们在那里有班级和次类。 个人之间有着二进制的关系。 00:03:52.400 EN: These are called properties. Or when you use the logical term, also roles. Properties also can be applied to relations. 中文初译: 这些属性被称为属性。 或者当您使用逻辑术语时, 也可以使用角色。 属性也可以适用于关系 。 00:04:04.590 EN: So these relationships between instances, those properties can have certain properties themselves, for example, domain and range, which we have already discussed in RDF schema. 中文初译: 因此,这些实例之间的关系, 这些属性本身可以有某些属性, 例如域和范围, 我们已经在“RDF”模型中讨论过。 00:04:18.070 EN: But other properties are, for example, transitive or inverse functional and inverse functional, which we will discuss today. 中文初译: 但其他属性则是过渡性的或反向的功能和反向功能,我们今天将讨论这些特性。 00:04:32.320 EN: There are also data types, for example, numbers, dates, strings, which are called in the logical world concrete domains. 中文初译: 还有一些数据类型,例如数字、日期、字符串等,在符合逻辑的世界具体域中被称作数据类型。 00:04:41.970 EN: There are logical expressions which can be used. And ontologies have a clear semantics, which can be also formally defined similarly as we have defined the semantics of RDF schema in the last lecture. 中文初译: 有一些逻辑表达式可以使用。 肿瘤有清晰的语义, 也可以正式定义, 就像我们在上次演讲中定义了“ RDF” 的语义一样。 00:05:01.020 EN: So for ontologies, there is a standard called OWL, the Web Ontology Language. So OWL is an acronym for Web Ontology Language. 中文初译: 所以对本体论来说,有一个叫做“OWL”的标准,即网络本体论语言。所以“OWL”是网络本体论语言的缩略语。 00:05:12.860 EN: And since it can be more easily pronounced, it's called OWL and not vowel. It's actually a family of languages for authoring ontologies. 中文初译: 因为它可以更容易地被宣告, 它被称为“ OWL ” 而不是元音。 它实际上是一个语言的家族, 用于撰写本体论。 00:05:24.380 EN: Since 2004, it's a W3C recommendation, so a standard, official standard. And five years later in 2009, the second revision of OWL was standardized, which is still in effect. 中文初译: 自2004年以来,这是一个W3C建议,因此是标准的官方标准。五年后,在2009年,“OWL”第二次修订实现了标准化,目前仍然有效。 00:05:40.400 EN: It's a semantic fragment of first order logic. So FOL stands for first order logic. Most of you have heard that probably in a logic's lecture. 中文初译: 这是一阶逻辑的语义片段。 所以FOL代表一阶逻辑。 你们中的大多数都听说过, 可能是在逻辑的演讲中听到的。 00:05:54.660 EN: And the W3C documents, they contain more details, of course, so when you are looking up the standard on the W3C web page, you will find more detailed information about the things. 中文初译: W3C文件包含更多细节, 当然, 当你在 W3C 网页上寻找标准时, 你会找到关于这些事情的更详细信息。 00:06:11.520 EN: So we can already use RDF schema as a simple ontology language. RDF schema is suitable for simple ontologies. The advantage is that the inferencing is relatively efficient. 中文初译: 因此,我们已经可以使用“RDF”模型作为简单的本体论语言。“RDF”模型适合简单的本体论。其优点是推论相对有效。 00:06:26.280 EN: So we have seen that last week, or actually two weeks ago, that we can apply these inference rules and we can compute and infer new statements, new triples, and add them to our knowledge base. 中文初译: 因此,我们上周或实际上两周前看到,我们可以应用这些推论规则,我们可以计算和推断新的声明,新的三重声明,并将其加入我们的知识库。 00:06:43.120 EN: But it's not suitable for more complex modelings. There are more powerful languages. One of them is OWL, but also F logic, for example, and then logics in general. 中文初译: 但它不适合更复杂的模型。 有更强大的语言。 其中之一是“ OWL ” , 但也包括 F 逻辑, 以及一般的逻辑。 00:06:59.560 EN: There are a zoo of different logics, which add more complexity and more expressivity basically to knowledge representation. 中文初译: 有一种由不同逻辑构成的动物园,这基本上增加了知识代表性的复杂性和表达性。 00:07:10.180 EN: OWL in a way tries to be a sweet spot between on the one hand being manageable, on the other hand also provide some expressivity. 中文初译: “OWL” 试图成为一个温馨的地方, 一方面可以管理, 另一方面也可以提供一些表达性。 00:07:21.020 EN: There's also a special syntax for OWL, which is called Manchester syntax, which is a bit more compact than the turtle syntax. 中文初译: 曼彻斯特语法,比海龟语法要紧一点。 00:07:30.880 EN: For example, in turtle, we would write Einstein RDF type professor. And in the Manchester syntax, we use the keywords individual and types to indicate that Einstein is an individual and it belongs to the type professor. 中文初译: 例如,在海龟中,我们会写爱因斯坦“RDF”类教授。而在曼彻斯特语法中,我们用关键词个人和类型来表示爱因斯坦是个体,属于该类教授。 00:07:50.120 EN: So that's a bit different. There's a bit of different syntax, but you can also use turtle syntax. So all OWL ontologies can be expressed in turtle, but the logicians have created a bit their own syntax, which they prefer to turtle, but it doesn't really matter which one to use. 中文初译: 所以这有点不同。有些语法不同, 但你也可以使用海龟语法。所以所有的“OWL”内涵都可以用海龟表达, 但是逻辑学家创造了他们自己喜欢的词法, 他们喜欢海龟, 但用哪一种并不重要。 00:08:11.840 EN: So an OWL document, of course, is also an RDF document. That's why we have these prefixes in the header of the document. 中文初译: 所以“OWL”文件当然也是“RDF”文件。这就是为什么我们把这些前缀放在文件页眉中的原因。 00:08:21.660 EN: We can define namespace prefixes in order to make the body of the document more readable. And then there can be some general information in the header of the OWL document. 中文初译: 我们可以定义命名空间前缀, 以使文档的正文更容易读取。 然后在“ OWL” 文档的页眉中可以有一些一般信息 。 00:08:38.990 EN: So you see this here. For example, we have a bookstore ontology. We can say that this is of type OWL ontology. So there's a special class for ontologies and then we can attach some meta information about this ontology. 中文初译: 所以请看这里。例如,我们有一个书店本体论。我们可以说这是“OWL”本体论类型。所以有专为本体论开设的特别班级, 然后我们可以附上一些关于这个本体论的元信息。 00:08:56.270 EN: For example, a comment from the RDF schema namespace. And here we can add in natural language that this is a bookstore ontology. 中文初译: 例如,“RDF” schema 命名空间的评论。在这里我们可以用自然语言补充, 这是一家书店的本体论。 00:09:05.260 EN: And then we have meta information, for example, about the version of the ontology. This version is 1.2 and it imports another ontology. 中文初译: 然后我们有了关于本体论版本的元信息。这个版本是1.2, 它又输入了另一种本体论。 00:09:14.890 EN: And then there is a URI given or URL where another ontology can be integrated or imported in the actual ontology. 中文初译: 然后有一个“URI” 或 URL , 可以将另一个本体论整合或导入到实际本体论中。 00:09:27.970 EN: So that makes ontologies also modular. You can have basically a network of ontologies which reference each other using this OWL imports statements. 中文初译: 这使得本体论也模块化。基本上你可以有一个本体论网络, 使用这个“ OWL” 进口报表互相参照。 00:09:41.390 EN: And as you can see here, these are all triples. Subject, predicate, object. So all OWL documents are basically sets of triples. 中文初译: 正如你在这里所看到的,这些都是三重的。 主题, 谓词, 对象。 所以所有“ OWL” 文件基本上都是三重的。 00:09:50.870 EN: We have here the semicolon. That means the subject is reused. So that's why the second, third, fourth triple here also use bookstore ontology as subject. 中文初译: 我们这里有分号。 这意味着主题被再利用了。 这就是为什么第二,第三,第四三 这里也使用书店本体论作为主题。 00:10:03.310 EN: And it ends with a full stop. So this is the turtle syntax for some metadata in an OWL ontology. So of course we can reuse all properties of which were defined, all resources which were defined in RDF schema. 中文初译: 最后是完全停止。这就是“OWL”肿瘤学中某些元数据的海龟语法。因此,我们当然可以重新使用所有被定义的属性,所有资源都是在“RDF”模型中定义的。 00:10:25.960 EN: In particular, RDFS comment, RDFS label, RDFS C also, RDFS is defined by, but OWL adds further information, further properties for versioning. 中文初译: 特别是“RDFS”评论,“RDFS”标签,“RDFS”C,“RDFS” C也由“RDFS”定义,但“OWL”增加了进一步的信息,为版本提供了进一步的属性。 00:10:40.780 EN: Like for example, OWL version info or OWL prior version to refer to a prior version of an ontology to indicate to which versions of ontologies it's backward compatible with that can be done using this property backward compatible with. 中文初译: 例如,“ OWL” 版本信息或“ OWL” 先前版本是指先前版本的本体论, 以表明它与哪些版本的本体论与使用此属性后向兼容的本体论相符。 00:11:01.260 EN: You can also indicate that a version of the ontology is incompatible with another version are using this incompatible with property. 中文初译: 您也可以指出,本体论的版本与另一个版本不兼容,使用此版本与属性不兼容。 00:11:11.000 EN: And then there are also the properties depreciated class and depreciated, depreciated, deprecated property, which basically indicate that certain classes and properties should not be used anymore because maybe you have defined a new version of it. 中文初译: 然后还有财产折旧类别 折旧类别,折旧,折旧,折旧的财产, 这基本上表明某些类别和财产不再使用 因为也许你已经定义了它的新版本。 00:11:32.680 EN: And that supports the evolution of ontologies because usually ontologies are never complete. Ontologies, they model the world, but the world changes, evolves. 中文初译: 这支持了本体论的进化,因为本体论通常从来就不完整。本体论是世界的模型,但世界的变化却在演变。 00:11:44.340 EN: Also the world of course is as complex as it can be. And ontologies are always a simplification and that's why they never complete in a way. 中文初译: 当然,世界也尽可能地复杂。本体论总是一个简单的简化, 这就是为什么它们从来就没有以一种方式完成。 00:11:56.040 EN: And they have to evolve, they have to change over time. And that's why it's important to indicate all also this compatibility or indicate when new definitions of classes or properties supersedes previous ones. 中文初译: 它们必须进化,它们必须随着时间的变化而变化。这就是为什么必须同时指出所有这些兼容性,或者指出当新的类别或属性定义取代了以前的类别或属性定义时。 00:12:12.260 EN: And then the previous definitions should be marked as deprecated. And then of course OWL imports, which I already explained, which can be used to import other files, ontology files and make them virtually part of the current file. 中文初译: 然后之前的定义应该被标记为折旧。 当然,“ OWL” 导入, 我已经解释过了, 这可以用来输入其他文件, 肿瘤文件, 并使它们成为当前文件的一部分。 00:12:33.120 EN: So OWL documents consist of sets of axioms. Axioms are usually expressed as sets of RDF triples. So an OWL document is basically also a set of RDF triples, nothing else. 中文初译: 所以“ OWL” 文件由数组轴组成。 轴通常以三组“ RDF” 表示。 因此, “ OWL” 文档基本上也是一组“ RDF” 三组, 仅此而已 。 00:12:47.100 EN: You can use RDF XML as the standard syntax, but also turtle is possible. So any RDF serialization can be also used for OWL ontologies. 中文初译: 您可以使用“ RDF” “ XML ” 作为标准语法, 但是海龟也可以使用。 所以任何“ RDF” 序列也可以用于“ OWL” 肿瘤学 。 00:13:01.220 EN: And in particular, turtle and some other formats are easier to read and to process. The syntaxes for OWL depend also a bit on the use case. 中文初译: 特别是,海龟和其他一些格式更容易阅读和处理。“OWL”的语法也取决于使用大小写。 00:13:12.780 EN: So RDF XML is the one which was standardized in the very beginning. There's also a special OWL XML version which is a bit easier to read for XML tools, which is a bit more simpler, but it's not 100% compatible with RDF. 中文初译: 所以“ RADF” “ XML” 是最初标准化的。 还有一种特殊的“ OWL ” “ XML ” 版本, 它比较容易读到“ XML” 工具, 更简单一点, 但与“ RDF ” 不完全兼容 。 00:13:34.460 EN: And then we have turtle. So that makes it very easy to read and write because we basically used also human readable version of an RDF serialization. 中文初译: 然后我们又有了乌龟。这样就很容易读写了, 因为我们基本上也使用人类可读的“RDF”序列化版本。 00:13:47.480 EN: And then there are some other syntaxes, but I would say most of the ontologies use RDF XML or turtle to be represented. 中文初译: 还有一些其他的语法, 但我想说大部分的本体论都使用“RDF”或“XML”来代表。 00:13:58.880 EN: So RDF XML usually, since it's the most standardized serialization of RDF, is often the duty for publication that you have an RDF XML version, and then you can provide other serializations optional of your ontology. 中文初译: 所以"RDF""XML"通常, 因为它是最标准化的 "RDF"序列化, 经常是发布的义务, 你有“RDF”的“XML”版本, 然后您可以提供其他的序列化 来选择你的本体论。 00:14:16.660 EN: And all files consists of a header with general information and then the rest with the actual ontology. So some of this header information we have already looked at. 中文初译: 所有文件都包含一个带有一般信息的页眉, 其余的文件则包含实际的本体论。所以有些我们已经查看过的页眉信息。 00:14:30.040 EN: Here you can see a small example. So we have the prefix definitions on top of our document. We have then a statement indicating that this is an ontology here, this particular document. 中文初译: 您可以在这里看到一个小的例子。 因此我们的文件上面有前缀定义。 然后我们有一个声明, 表明这是一个本体论, 这个特定的文件。 00:14:47.140 EN: And then inside in the body, we have defined two classes. Persons, which is of type OWL class. So there is a new class or a new like identifier for OWL classes, which is very similar to RDFS class, which we have already seen a few weeks ago. 中文初译: 在身体里,我们定义了两类。 一类是“ OWL” 类的人。 因此, “ OWL” 类有一个新的类或新的类似标识符, 这与“ RDFS” 类非常相似, 我们几周前已经见过了。 00:15:13.420 EN: And student is also defined as OWL class. And then a student is defined as a subclass of a person here. So there the RDF schema property subclass of is reused. 中文初译: 学生也被定义为“OWL”类。然后,学生被定义为这里一个人的子类。因此,在这里,“RDF”的化学属性子类被重新使用。 00:15:28.320 EN: So OWL, of course, reuses the terms, the properties, the classes, which were already defined previously in RDF and RDF schema. 中文初译: 所以“OWL”当然会重新使用 先前在“RDF”和“RDF”模式中定义的术语、属性、类别。 00:15:38.300 EN: So they nicely built on top of each other. And the separation between RDF, RDF schema, and OWL is also not always 100% clear because they developed historically a bit. 中文初译: 所以它们相互搭建的很好。“RDF”、“RDF”schema和“OWL”之间的分离 也并不总是百分之百清晰,因为它们在历史上有所发展。 00:15:50.240 EN: There was first RDF, then RDF schema, later OWL. And that's why some things are already defined in RDF schema, which are then reused later in OWL and vice versa. 中文初译: 起初是“RDF”,后来是“RDF”schema,后来是“OWL”。这就是为什么有些东西已经在“RDF”schema中定义了,然后在“OWL”中重新使用,反之亦然。 00:16:06.120 EN: So it's relatively concise and easy to read using turtle. It's not specifically designed for OWL and complex expression look a bit impractical. 中文初译: 所以它比较简洁,很容易用乌龟阅读。 它不是专门为“OWL”设计的, 复杂的表达方式看起来有点不切实际。 00:16:19.080 EN: We will see that, but there is very good tool support between turtle is also meanwhile in addition to RDF XML supported by almost all tools. 中文初译: 我们可以看到这一点,但除了几乎所有工具都支持的“RDF”“XML”之外,海龟之间也有非常好的工具支持。 00:16:32.860 EN: Then there's this Manchester syntax by Manchester because the University of Manchester was an important host of a research group working on description logics, which is the basis for OWL. 中文初译: 然后是曼彻斯特的曼彻斯特语法, 因为曼彻斯特大学是一个研究小组的重要东道主 研究描述逻辑, 这是“OWL”的基础。 00:16:49.120 EN: And they developed the syntax for OWL, which then is called Manchester syntax. So there, it looks a bit differently. 中文初译: 他们开发了“OwL”的语法,这叫曼彻斯特语法。所以看起来有点不同。 00:16:59.060 EN: We only showed this to you here for information. You don't have to write Manchester syntax yourself, but if you see it somewhere that you can understand and read it. 中文初译: 我们在这里只给你们展示了这个信息。你不必自己写曼彻斯特语法, 但如果你看到它的地方你可以理解和阅读它。 00:17:11.680 EN: So you also have a prefix definitions on top and then basically ontology colon and then the instance which belongs to the ontology or we have done a class colon looks a bit like maybe Java code where you declare variables, similar inspired from that is maybe a Manchester syntax. 中文初译: 因此,你在上面还有一个前缀定义, 基本上就是本体结肠, 然后是属于本体结肠或我们做过类结肠的例子, 看上去有点像爪哇代码, 你在这里宣布变量, 从中可以发现类似的变量, 也许曼彻斯特语法。 00:17:37.080 EN: So we have a class and this class is person and then it's defined to be a subclass of OWL thing. And similarly a class student and it's defined here to be a subclass of a person. 中文初译: 所以,我们有一个班级,这个班级是人, 然后它被定义为“OWL”的子班。同样,一个班级学生, 也在这里被定义为一个人的子班。 00:17:52.920 EN: You can see here OWL thing. OWL thing is a class of everything, the class which contains everything. So it's a superclass of everything. 中文初译: 你可以看到这里的“OWL” 东西。“OWL”是所有事物的一类, 包含所有事物的类。所以它是一个所有事物的超级类。 00:18:05.360 EN: That's why this statement is in a way redundant. By default, all classes are subclasses of OWL thing, but it also doesn't hurt to add it specifically. 中文初译: 这就是为什么这个语句在某种意义上是多余的。默认情况下,所有类别都是“OWL”小类, 但具体添加它也无妨。 00:18:17.680 EN: So the Manchester syntax is also easy to read and write. Some types of axiom also a bit difficult here. It follows a bit more description logic syntax. 中文初译: 因此曼彻斯特语法也很容易读写。 这里有些类的正弦也有点困难。 它沿循了略微多一点描述逻辑语法。 00:18:32.740 EN: For example, you would write that student is a subclass of person using the set notation here like a subset because that's also how you can imagine a classes as subsets and inspired by this description logic syntax. 中文初译: 例如,你会写下,学生是个人的一个子类 使用设置的符号在这里像子集一样, 因为这也是你如何想象一个类作为子集 并受到这个描述逻辑语法的启发。 00:18:54.800 EN: The Manchester syntax evolved. It's a functional syntax and it's for example, used for the open biomedical ontologies, which is one quite famous, important renowned set of ontologies used in the life science area. 中文初译: 曼彻斯特语法进化了。这是一个功能性语法,例如,它用于开放生物医学理论学, 这是一套在生命科学领域使用的非常有名、重要的著名理论学。 00:19:12.860 EN: And then of course we can use RDF XML. Here you see the same example in RDF XML. So we have ontology defined here and in the ontology we have two classes. 中文初译: 当然,我们可以使用“RDF”“XML”。这里您可以看到“RDF”“XML”中的相同例子。所以我们在这里和在肿瘤学中定义了本体论,我们分两个班。 00:19:27.200 EN: The first class person, we use this RDF about attribute for this XML tag all class. And then we can indicate that the subclass of all thing. 中文初译: 头等舱的人, 我们使用这个“ RDF” 来描述这个 XML 标签的属性。 然后我们就可以显示所有事物的子类 。 00:19:42.720 EN: Here you see that you have to use the property entity references when you use namespaces inside attribute values here. 中文初译: 您可以在此看到, 当您在这里使用属性值内的命名空间时, 您必须使用属性实体引用 。 00:19:54.760 EN: And then we have defined the second class student and the subclass of person. So this of course also has very good tool support, but it's quite verbose and a bit difficult to read for humans, especially when ontologies become larger and more complex. 中文初译: 然后我们定义了二等学生和次等学生。所以这当然也有非常好的工具支持, 但对于人类来说,这是相当多的动词和有点难读, 特别是当本科学变得更大、更复杂的时候。 00:20:13.880 EN: Then there is this OWL XML syntax, which we also show you here for information. It's part of the W3C recommendation and it was specifically designed for OWL. 中文初译: 还有这个“OWL”“XML”语法, 我们在此也向您展示此语法以供参考。 这是 W3C 建议的一部分, 专门为“ OWL” 设计的 。 00:20:27.940 EN: So it's more simple than RDF XML, but still relatively verbose. Yeah, and I would prefer or generally I think most people prefer RDF based syntaxes. 中文初译: 所以比"RDF""XML更简单,但还是相对的动词。是的,我更喜欢或者一般地说,我想大多数人都喜欢基于"RDF"的语法。 00:20:43.680 EN: So this OWL XML syntax, I would say is not very common, but once you encounter it, here's how it looks like. You have a tag for ontology, then you can add a URI of the ontology as an attribute here. 中文初译: 所以这个"OWL""XML"语法, 我想说并不常见, 但是一旦你遇到它, 这就是它的外观。 你有肿瘤学标签, 然后你可以在这里添加一个“ URI” 作为本体论的属性 。 00:20:58.400 EN: Then you define a subclass of, and then you have basically two classes. The first one is person, the second is thing here. 中文初译: 然后定义一个子类,然后基本上分为两类。第一个是人,第二个是这里的东西。 00:21:06.660 EN: So it defines things using the properties as tags and indicating the subject and the object inside the tag subclass of and the same here for student, which is subclass of person. 中文初译: 所以它定义了用属性作为标签的东西, 并标出对象和对象 在标签子类内 00:21:26.720 EN: So ontologies contain three components or axioms, which are in a way analogous to RDF schema. So we have individuals or objects in RDF schema. 中文初译: 因此,在“RDF”的模型中, 原子学包含三个成分或轴, 与“ RDF” 模型类似。 所以我们在“ RDF” 模型中有个人或物体。 00:21:42.160 EN: We call them also instances. So these are concrete elements in a modeled world. Then we have classes or concepts, which are sets of objects and then roles or properties in RDF, which associate two individuals. 中文初译: 我们还称之为实例。因此,这些是模型世界中的具体元素。然后我们有分类或概念,它们是一组物体,然后是“RDF”中的角色或属性,它联系了两个人。 00:22:05.160 EN: So classes are basically the same as RDF schema classes. So we can, for example, say that professor is of type, our class, and in Manchester syntax, we would just write class colon professor. 中文初译: 所以班级基本上和“RDF”的样板课一样。比如说,我们可以说教授是类型,我们的班级, 在曼彻斯特语法中,我们只写一个类结肠教授。 00:22:25.260 EN: And as I already mentioned, we have two predefined classes, our thing, which is the superclass of everything. So everything is a subclass of our thing. 中文初译: 正如我已经提到过的,我们有两个预设的班级, 我们的东西,就是一切的超级班级。所以每件事都是我们事物的一个子班。 00:22:35.540 EN: All classes are subclasses of our thing. And that means that all instances, all objects, all entities are also instances of our thing. 中文初译: 所有类别都是我们事物的子类。 这意味着所有实例,所有物体,所有实体 也都是我们事物的例子。 00:22:46.360 EN: So all classes are subclasses of our thing and all instances are also instances of our thing. And then we have all nothing. 中文初译: 所以所有班级都是我们事物的子类, 所有的例子也是我们事物的实例。然后我们一无所有。 00:22:55.600 EN: And you can already figure out that nothing is the empty class. So it's a subclass of all classes. All nothing can be defined as a subclass of all classes because it's the empty class, corresponds to an empty set, right? 中文初译: 而且您已经可以发现,没有哪个是空类。 所以它是所有类的子类。 没有什么可以被定义为所有类的子类, 因为它是空类, 对应的是空组, 对不对? 00:23:11.400 EN: And an empty set is a subset of all other sets. That's why our nothing is a subclass of all classes. And it doesn't contain any instances. That's very important. 中文初译: 空套件是所有其他套件的子集。 这就是为什么我们没有任何东西 是所有分类的子类。 它并不包含任何实例。 这是非常重要的。 00:23:26.590 EN: So then we have abstract roles and we show you how they are defined in turtle here. They are defined like classes. 中文初译: 所以我们有抽象的角色, 并展示它们是如何在海龟中定义的。 它们像类一样定义的。 00:23:36.990 EN: Maybe you remember we could define properties in RDF schema, but in RDF schema, we have not distinguished properties which have objects, URIs as object or data values as objects. These were all properties. 中文初译: 也许你记得,我们可以在“RDF”模型中定义属性,但在“RDF”模型中,我们没有区分有物体的属性,“URIs”作为对象或数据值。这些都是所有属性。 00:23:56.730 EN: In OWL, we distinguish two types of properties, the properties which have as objects instances or other objects. 中文初译: 在“OWL”中,我们区分两类属性,即作为物体实例或其他物体的属性。 00:24:07.770 EN: These are called object properties. And then there are also data type properties which we come to on the next slide. 中文初译: 这些被称为对象属性。然后还有数据类型属性,我们在下一张幻灯片上找到这些属性。 00:24:17.690 EN: So here we have defined the property belongs to and it's of type OWL object property. That means the values, the objects of this property should be instances, resources, URIs. 中文初译: 因此,我们在这里定义了财产的属性和“OWL”类型的物体属性。这意味着价值,这些财产的物件应该是实例、资源、“URIs”。 00:24:37.250 EN: And then we can of course use domain and range definitions for them as well. So we can, for example, add that the domain of belongs to is a person and the range is organization. 中文初译: 然后,我们当然也可以为他们使用域和范围定义。例如,我们可以添加, 属于一个人的域, 而范围就是组织。 00:24:53.920 EN: So here we use what was defined already in RDF schema. In Manchester syntax, it would like that. We say object property colon and then the object property and domain and range can be indicated by simply stating domain colon, person, range colon, organization, that's how it looks in Manchester syntax. 中文初译: 因此在这里我们使用“ RDF” 语法中已经定义的内容。 在曼彻斯特语法中, 它希望这样。 我们说客体属性结肠, 然后用简单表达域号、 人、 范围结肠、 组织来表示对象属性和域域和范围, 这就是曼彻斯特语语法中的外观。 00:25:22.380 EN: Then we have concrete roles. And these concrete roles are now properties which have a data type in the range. Data, they're also called data type properties. 中文初译: 然后我们有了具体的角色。这些具体的角色现在都是属性, 在范围中具有数据类型。数据,它们也被称为数据类型属性。 00:25:34.720 EN: And that's why we can write here data type property first name in Manchester syntax, for example, and then define domain and range. 中文初译: 这就是为什么我们可以在这里写入数据类型属性名, 比如在曼彻斯特语法中, 然后定义域和范围。 00:25:44.120 EN: So the domain is a class person in that case, the range is a data type, in that case your string. So that means that the value of these concrete roles of these data type properties need to have data types as, yeah, range. 中文初译: 域是类人, 如果是这样的话, 范围是一个数据类型, 在这样的情况下, 您的字符串。 这意味着这些数据类型属性的这些具体角色的价值 需要包含数据类型, 如, 是, 范围 。 00:26:03.260 EN: That means that in triples where you have this property as a predicate, in the predicate position, the object must be a data value, a literal. 中文初译: 这意味着,在以三倍数值表示的您将这种属性作为谓词时,在上游位置上,对象必须是数据值、字面意思。 00:26:18.840 EN: So now again, in turtle syntax, first name type, our data type property, and then we have defined here also domain and range. 中文初译: 现在,在海龟语法中, 首字母类型, 我们的数据类型属性, 然后我们在这里定义了域和范围。 00:26:34.230 EN: Domain, person, and range, and here we indicate XSD string, we use XML schema data type as a range. Then we have individuals, individuals or instances. 中文初译: 域、 人和范围, 在此我们表示 XSD 字符串, 我们使用“ XML” 化学模型数据类型作为范围。 然后有个人、 个人或实例 。 00:26:51.390 EN: So for example, we can say that Einstein is of type professor. And then we have the property belongs to, and we have Princeton and Byrne, and we have a first name, the first name, so should be intended here. 中文初译: 比如说,我们可以说爱因斯坦是类型的教授。然后我们拥有属于我们的财产, 我们拥有普林斯顿和伯恩, 我们有一个名字,一个名字,这个名字,所以应该在这里使用。 00:27:16.440 EN: But since we have a semicolon, means that also for the last triple, the same subject applies. So there's also Einstein has first name, and since this is a data type property, a concrete role, we now use a data value, a literal, and we write Albert and use XSD string to take this kind of data value, this literal. 中文初译: 但是,既然我们有一个分号, 意思是最后三个主题也适用同一主题。 因此还有爱因斯坦有一个名字, 因为这是一个数据类型属性, 一个具体的角色, 我们现在使用一个数据值, 一个字面的, 我们写艾伯特, 用 XSD 字符串来取出这种数据值, 这个字面的。 00:27:45.470 EN: So roles are generally not functional. What does it mean? Functional means functional is something which can be used only once, yeah? 中文初译: 所以角色一般不起作用。这意味着什么?功能意味着功能是只能使用一次的东西,是吗? 00:27:56.100 EN: So roles are by default not functional. You can use, for example, first name, you can use as often as you want if someone has several first names. 中文初译: 所以角色默认为非功能性。 例如, 您可以使用名, 如果某人有几个名字, 可以随心所欲地经常使用名 。 00:28:05.240 EN: Or if you have Mary 2, you can use the property Mary 2 as many times as you want. By default, properties can be used with one individual, one instance, as many times as you like. 中文初译: 或者,如果您拥有玛丽2号,您可以使用玛丽2号地产的两倍。默认情况下,属性可以使用一个个体,一个实例,也可以使用多少次。 00:28:22.260 EN: We will see a bit later how we can make them functional. And here is, again, in Manchester syntax, how it looks like. 中文初译: 我们稍后会看到如何让它们发挥作用。这里是,在曼彻斯特语法中,它看起来如何。 00:28:31.580 EN: We say individual colon Einstein. We indicate the type, that it belongs to the type person, and then we have some facts, belongs to Princeton, belongs to Byrne, and first name Albert. 中文初译: 我们说个别的结肠爱因斯坦。我们指出的类型, 它属于类型的人, 然后我们有一些事实, 属于普林斯顿,属于伯恩, 名字是阿尔伯特。 00:28:52.290 EN: So here we have a UML diagram. For those of you who are familiar with UML, how classes can be organized in OWL. 中文初译: 我们这里有一个 UML 图表。 对于熟悉 UML 的人来说, 如何在“ OWL” 中组织课程 。 00:29:07.210 EN: So we have class axioms, and there are different types of class axioms, and class axioms are a special type of axioms. 中文初译: 所以,我们有类轴, 并且有不同种类的类轴, 类轴是一种特殊的类轴。 00:29:20.550 EN: So subclass of, for example, is a class axiom. Equivalent class is a class axiom. Disjoint classes, disjoint union are also class axioms. 中文初译: 例如,子类是类的轴。 等值类是类的轴。 等值类是类的轴。 分解类, 离合也是类的轴。 00:29:31.830 EN: And then there are class expressions, which can basically be used to build new classes or we can define classes as nominals, basically, or define classes with a name, named classes, basically. 中文初译: 然后是类表达式, 基本上可以用来构建新分类, 或者我们可以将类别定义为标称的, 基本,或者以名称命名的分类来定义。 00:29:56.850 EN: So that's how the structure of these different class relations look like. We will now look at some examples, how this looks like. 中文初译: 所以这些不同阶级关系的结构是这样的。我们现在将研究一些例子, 如何看待这些例子。 00:30:07.650 EN: One of these class axioms is our equivalent class. And equivalent class basically means that two classes are equivalent. 中文初译: 其中一个类轴就是我们的等同等级。而同等等级基本上意味着两个类是等的。 00:30:20.900 EN: So let's look maybe at this example here. In turtle, we have book as of type class. It's a subclass of publications. Also publication is a class. 中文初译: 让我们来看看这个例子。在海龟中,我们有书作为类类。它是出版物的子类。出版也是类类。 00:30:38.030 EN: And publication has the equivalent class publications. So please note the S here. Let's assume you have defined two classes, publication and publications. You can make them equivalent. 中文初译: 出版品有等同的类出版物。请在此注意S。假设你已经定义了两类, 出版品和出版物。你可以把它们等同起来。 00:30:54.230 EN: And then also publications is defined as our class. For example, this sometimes happens. So if you develop an ontology, you should think about, before you start developing your ontology, how you want to name classes. 中文初译: 出版也被定义为我们的类。例如,有时会发生这种情况。所以如果你发展一种本体论, 你应该在开始开发本体论之前考虑一下, 如何命名课程。 00:31:10.210 EN: Usually you name classes either in singular or plural, right? But sometimes you might make a mistake. If you didn't think about this in the beginning, you might define some classes in singular, some classes in plural. 中文初译: 通常您会用单数或复数来命名类, 对吧? 但有时你可能会犯错。 如果您在开始时不考虑这一点, 您可能会用单数来定义某些类, 有些以复数来定义。 00:31:24.010 EN: And then you want to then later on, in the next version of your ontology, define everything in singular, for example. 中文初译: 然后你想在以后, 在下一个版本的你的本体论中, 以单数来定义一切,例如。 00:31:31.730 EN: Then you have to introduce a new class publication. And then you might want to define them as equivalent class to the previously defined publications. 中文初译: 然后,你必须引入一个新的分类出版物。然后,你也许应该把它们定义为与以前定义的出版物相当的类别。 00:31:43.070 EN: And of course you would then indicate that publications is deprecated so that it should not be used any more further. 中文初译: 当然,你然后会指出,出版物已经过时,因此不再使用。 00:31:50.890 EN: But by making them equivalent, you basically make it backward compatible. And on the button here for information, it's also indicated how it looks in the Manchester syntax for your information. 中文初译: 但是通过把它们等同, 你基本上可以使它向后兼容。 在此处用于信息的按钮上, 它也表明了它在曼彻斯特语法中如何看待你的信息。 00:32:09.070 EN: But as I said, Manchester syntax, we just added for you as information will not be part of the exam later on. So subclass of is basically the same as an RDF schema. 中文初译: 但正如我所说,曼彻斯特语法,我们只是为你补充了信息,因为信息不会成为以后考试的一部分。所以子类基本上和“RDF”计划一样。 00:32:29.220 EN: So we can define classes to be subclasses of each other. For example, defining a faculty staff to be a subclass of person as here in this example, and professor being a subclass of faculty staff. 中文初译: 这样我们可以将班级定义为彼此的子类。例如,将教师定义为本例中的子类人员,将教授定义为教师的子类人员。 00:32:45.440 EN: And with this, you see that we can create a hierarchy. So you have professor, subclass of faculty staff, and faculty staff, subclass of person. 中文初译: 通过这个,你可以看到我们可以建立一个等级制度。所以你有教授, 教职员工的子级, 教师的子级, 人类的子级。 00:32:54.260 EN: So you can create large hierarchies of taxonomies of classes, class taxonomies. And that is something which appears quite often in ontologies that we create such hierarchies, taxonomies of classes. For example, in species, it appears a lot. 中文初译: 所以,你可以建立大型的等级分类, 包括班级分类, 班级分类。这是在肿瘤学中经常出现的东西, 我们常常会创建这样的等级分类, 班级分类。例如,在物种方面,它看起来很多。 00:33:16.480 EN: Of course, there are different types of persons, different types of cars. So this classification of things is one very important aspect of knowledge organization, of knowledge management. 中文初译: 当然,有不同类型的人和不同的汽车。因此,对事物的分类是知识组织、知识管理的一个非常重要的方面。 00:33:30.900 EN: Then there is our disjoint with. So previously we could define two classes. We could define car and we could define person, but no one was preventing us from making an instance, an instance of two classes. 中文初译: 然后,我们又与它脱节了。所以以前我们可以定义两个类别。我们可以定义汽车,我们可以定义人,但是没有人阻止我们举一个例子,一个是两个类别的例子。 00:33:49.300 EN: Yeah, we could define zone R as an instance of person and zone R as an instance of car, which doesn't make sense. 中文初译: 是的,我们可以把R区定义为 人的例子,R区定义为汽车的例子, 这没有意义。 00:33:57.660 EN: And that's why OWL introduces the property disjoint with. This is an object property or an abstract role and the logic terminology. 中文初译: 这就是为什么“OWL” 引入了与财产脱节的属性。 这是物体属性, 或者抽象的角色和逻辑术语。 00:34:08.980 EN: And that allows us to define two classes of being disjoint. Disjoint means there are no instances which belong to both classes. 中文初译: 这让我们可以定义两种分解的类别。 分解意味着不存在属于这两种类别的情况。 00:34:19.320 EN: The same as in set theory. You have two disjoint sets and similarly you have two disjoint classes. So here in this example, we have book defined as a class. 中文初译: 与设定理论相同。 您有两个脱节组, 类似地您有两个脱节类。 因此在此示例中, 我们的书被定义为一个类 。 00:34:33.020 EN: It's a subclass of publication. And then we have faculty staff is also a class. And then we define that faculty staff is disjoint with publication. 中文初译: 这是出版的子类。然后我们有教职员也是班级。 然后我们定义了教职员与出版脱节。 00:34:44.300 EN: Makes a lot of sense to define persons to be disjoint from publications. By default, this is not the case in RDF, RDF schema, OWL, we have this open world assumption. 中文初译: 定义与出版物脱节的人是有道理的。 默认情况下,在“RDF ” 、 “RDF”schema、“OWL ” 中情况并非如此。 我们有一个开放的世界假设。 00:35:01.980 EN: That means basically if something is not set to be not an instance, it could still be an instance. We just don't know it yet. 中文初译: 这基本上意味着,如果有些事情不是注定要成为一个实例,它可能仍然是一个例子。我们只是还不知道而已。 00:35:12.720 EN: That's we have learned in the RDF schema lecture two weeks ago, how this inferencing works. So with the inferencing, we can reveal implicit information. 中文初译: 这就是我们两周前在“RDF”计划演讲中学到的, 这个推论是如何起作用的。因此通过推论,我们可以揭示隐含的信息。 00:35:22.620 EN: And sometimes from this inference reasoning, it become, we can infer that something belongs to a class which was not explicitly stated. 中文初译: 有时从这种推论推理中,我们推断出 某种东西属于一个没有明确说明的类别。 00:35:36.900 EN: And that's why it's very important to state this disjoint with axioms so that we know which classes should be really disjoint. 中文初译: 这就是为什么说这种与正弦脱节 非常重要, 这样我们才能知道哪些类应该真的脱节。 00:35:48.740 EN: And then we have here in the example professor as a class and defined as a subclass of faculty staff and publication defined as a class. 中文初译: 然后我们在这里的例子中,教授作为一个班级 被定义为教师的子班级 出版被定义为一个班级。 00:35:58.580 EN: So this is already a small ontology defining books, publications, professors, faculty, staff. Then we have individuals and class relations. 中文初译: 所以这已经是一个小的本体论 定义书籍,出版物,教授,教师,员工。然后我们有了个人和阶级关系。 00:36:13.060 EN: So here in this example, we have a book defined, which belong, which is a book we have defined as type all class. 中文初译: 因此在此示例中, 我们定义了一本书, 属于它, 这是本书, 我们定义了所有类的类型。 00:36:24.720 EN: It's a subclass of publication. And then we have a concrete book defined here, semantic web kundlag, which is a book where basically this lecture is also based on. 中文初译: 这是出版的子类。然后我们有了一本具体的书, 这里定义了语义学的网页 kundlag, 这本书基本上也是这个讲座的基础。 00:36:37.220 EN: So this book is also recommendable for you. If you want to read a bit more details and get all the information, you can read this book, which is now defined as a book. 中文初译: 因此这本书对您也是推荐的。 如果您想要阅读更多细节并获取全部信息, 您可以读读这本书, 这本书现在被定义为一本书 。 00:36:50.020 EN: So it belongs to the class book. And then we have defined authors. And here in that case, we have defined four authors. 中文初译: 因此它属于班本。然后我们定义了作者。在这种情况下,我们定义了四个作者。 00:36:59.720 EN: Yeah, Marcus Krotch, Pascal Hitzler, Sebastian Rudolf and Jorg Zuhrer. So in Manchester syntax, it would like that. 中文初译: 马库斯·克罗奇、帕斯卡尔·希茨勒、塞巴斯蒂安·鲁道夫和约格·祖赫尔德。所以在曼彻斯特的语法中,它希望如此。 00:37:10.020 EN: And since we have defined book as a subclass of publication, it also follows that the book semantic web kundlag is also a publication, yeah. 中文初译: 既然我们把书定义为出版的子类, 由此也可以看出, 书的语义网 kundlag 也是一本出版物,是的。 00:37:21.720 EN: Instances of subclasses are also instances of superclasses. We discussed this already three weeks ago in the RDF schema lecture, but of course the same applies to OWL. 中文初译: 亚类的例子也是超级类的例子。我们三周前在“RDF”计划演讲中已经讨论了这个问题,但当然这也适用于“OWL”。 00:37:33.120 EN: And that's something very important to keep in mind when you build ontologies, sometimes it's a bit unclear what are instances, what are subclasses. 中文初译: 当你建立基础学时, 这一点非常重要。 有时候,有些事情还不清楚什么是实例,什么是子类。 00:37:41.760 EN: You always have to imagine subclasses as subsets, and then it becomes relatively clear what should be modeled as classes and what should be modeled as instances of classes, if you imagine this as sets and subsets. 中文初译: 你总是必须想象子类作为子类, 然后它变得相对清楚什么应该被建模为类, 什么应该被建模为类的例子, 如果你把它想象成组和子类的话。 00:37:58.520 EN: So, and then we have role relationships, so relationships between properties, analogous to classes. For roles, there is RDF as subproperty of and equivalent property, but roles can also be inverses, and that's called OWL inverse of. 中文初译: 所以,然后我们有角色关系, 属性之间的关系, 类似于类别。对于角色, 有“RDF”作为子财产和等同财产, 但角色也可以是反面的, 这就是所谓的“OWL ” 反面的。 00:38:20.420 EN: So for example, here we have internal syntax defined eximined by OWL inverse of examiner of, yeah. So these properties are defined to be inverses of each other. 中文初译: 例如,这里我们有内部语法定义 由“ OWL” 反之的检查者来解释, 是的。 因此这些属性被定义为彼此的反义 。 00:38:36.680 EN: What does it mean? It means basically when you have a subject, predicate and object, and the predicate is for example, eximmed by, you have Peter is eximmed by Paul, then you can basically turn the statement around and you can say, Paul is the examiner of Peter. 中文初译: 这意味着什么?它基本上意味着当你有一个主题, 上游和对象, 而上游是例如, 被挤出来, 你让彼得 被保罗挤出来, 然后你可以基本上把口供转过身来, 你可以说,保罗是 Peter的检查员。 00:38:56.380 EN: So you can infer new triples by just turning them around and replacing the property with the inverse property. So then there are also role properties, domain and range, we know already from RDF schema. 中文初译: 因此,您可以推导出新的三重, 只需翻转它们, 用反向属性替换财产。 然后还有角色属性、 域和范围, 我们已经从“ RDF” 模型中知道。 00:39:13.380 EN: Then there is transitivity, so we can define that the property role is transitive. What does it mean transitive? 中文初译: 然后是过渡性,这样我们就可以确定财产的作用是过渡性的。 这意味着什么是过渡性? 00:39:21.880 EN: So if we have AB in a relationship with this role, and we have BC in the relationship with this role, it implies that AC is in the relationship with this role. 中文初译: 所以如果我们有AB 与这个角色的关系, 而我们有BC与这个角色的关系, 这意味着AC和这个角色的关系。 00:39:33.960 EN: For example, subclass of is a transitive relationship and sub-property of. Also we have symmetry. So if we have, for example, AB in a relationship with each other, if the role R would be defined to be symmetric, it implies that also BA holds. 中文初译: 例如,亚类是一种过渡关系和次级财产。我们也有对称性。因此,如果我们在彼此的关系中有AB,如果R的作用被界定为对称性,它意味着BA也持有。 00:39:57.180 EN: Yeah, for example, being married to would be a symmetric role, right? If one person is married to another person, then this other person is also married to the first person, right? 中文初译: 比如说,结婚就是对称角色,对吗?如果一个人嫁给另一个人,那么这个人也嫁给第一个人,对吧? 00:40:10.540 EN: That's why being married to should be defined as symmetric. Then we have functionality. If we have AB in a relationship and AC in a relationship, it would imply that B and C are the same. 中文初译: 这就是为什么结婚应该被定义为对称。 然后我们就可以实现功能。 如果我们有AB在关系中, AC在关系中, 这意味着B和C是相同的。 00:40:27.660 EN: Do we have an example for that? And Germany also married to would apply here if A is married to B and A is married to C, it would imply that B and C, if it's legal, should be the same, should refer to the same real world person. 中文初译: 我们是否有这方面的例子?如果A与B和A与C结婚,德国也愿意在此申请,这意味着如果B和C合法,那么B和C应当相同,应该指同一个真实世界的人。 00:40:47.540 EN: Why do we have this? Because in RDF, RDF schema on the semantic web in general, we don't have a unique names assumption. 中文初译: 为什么我们有这个?因为一般语义网站上的“RDF ” “RDF”模式, 我们没有独特的名字假设。 00:40:54.320 EN: So we can have different identifiers for the same real world thing, yeah? So for one person, there might be different identifiers. 中文初译: 所以,我们可以有不同的识别特征 对于相同的真实世界的事情,是的。所以对于一个人来说,可能有不同的识别特征。 00:41:03.220 EN: For example, your employer might create an identifier for you. Your tax authorities might create an identifier for you. 中文初译: 例如,您的雇主可能会为您创建身份识别特征。 您的税务主管部门可能会为您创建身份识别特征。 00:41:12.000 EN: Your email provider has an identifier for you. So there are lots of identifiers for people. And that's why sometimes we can infer that two identifiers must refer to the same real world person. 中文初译: 您的电子邮件提供商有您的识别信息。 因此人们有很多识别信息。 所以我们有时可以推断, 两个识别信息必须针对同一个真实世界的人。 00:41:27.460 EN: And that's, for example, could be done by defining some property as being functional. And then we have also inverse functionality. 中文初译: 比如说,通过定义某些属性是功能性的,就可以做到这一点。然后我们也有反向的功能。 00:41:36.660 EN: So if we have AB in a relationship and we have CB in a relationship, this would imply that A and C are equal. So what could be an example for that? 中文初译: 因此,如果我们有AB关系和CB关系,这将意味着A和C是平等的。 那么什么可以作为这方面的范例呢? 00:41:49.380 EN: For example, has email, yeah? If a person A has the email B and the person C has the same email B, it would imply that the person A is the same as person C because usually an email refers to one particular person. 中文初译: 例如,有电子邮件吗?如果一个人A拥有电子邮件B,而那个人C拥有相同的电子邮件B,这意味着这个人A和人C是同一个人C,因为电子邮件通常是指某个特定的人。 00:42:08.020 EN: Or maybe email is not the best example. We can also choose passport IDs, yeah? If there are two persons having the same passport ID, those two persons are probably the same. 中文初译: 也许电子邮件不是最好的例子。 我们还可以选择护照身份证,是吗? 如果两个人有相同的护照身份证,那两个人可能是一样的。 00:42:23.860 EN: So it's actually one person, two identifiers for the same person, two URIs. So let's look at some examples. So here, for example, we have Fenrich as a professor and we have defined same as Professor Fenrich. 中文初译: 所以其实是一个人,两个身份识别特征 同一个人,两个“URI”。让我们来举几个例子。例如,我们有芬里希作为教授, 我们定义与芬里希教授相同。 00:42:44.860 EN: This is actually a same as relationship. So that means that these two resources really refer to the same real world entity. 中文初译: 这实际上与关系是一样的。这意味着这两种资源 真正指的就是同一个真实的世界实体。 00:42:54.920 EN: They are exactly the same. And that happens quite often that we have different identifiers for the same real world thing. 中文初译: 它们完全一样。这种情况经常发生,我们对于同一个现实世界有不同的识别特征。 00:43:08.660 EN: And here we could also then apply some inferencing and we could infer that Professor Fenrich is a professor because Fenrich is a professor since it's same as Professor Fenrich, all triples which apply to Fenrich would also apply then to Professor Fenrich. 中文初译: 在这里,我们也可以进行一些推论 我们可以推断 芬里希教授是一位教授 因为芬里希教授是一位教授 因为芬里希教授和芬里希教授一样 所有适用于芬里希的三联体 也同样适用于芬里希教授 00:43:36.560 EN: And similarly, as we have equivalence or the equality of instances using same as, all same as, it's a very important property. 中文初译: 类似地,因为我们有等同或同等的例子 使用相同的, 和一样的,它是一个非常重要的财产。 00:43:48.180 EN: So please all remember this all same as. We can also define the two identifiers are really referring to different real world entities and that's all different from. 中文初译: 所以请大家记住这一点 和这些一样。我们也可以定义这两个识别信息 真正是指不同的真实世界实体, 这与现实世界不同。 00:44:04.240 EN: So that means that two individuals are really different because since we have the open world assumption, just that two things have different names, it does not mean that they are different. 中文初译: 这意味着两个人真的不同, 因为因为我们有开放世界的假设, 只是两样东西有不同的名字, 并不表示它们是不同的。 00:44:24.710 EN: They still could be the same. We just don't know it yet. And then there's also the possibility to define a set of distinct different individuals. 中文初译: 他们可能还是一样。我们只是还不知道。然后,还有一种可能性 来定义一组不同的人。 00:44:40.950 EN: In Manchester syntax looks like that. You can say different individuals, Jorg Zuhrer, Pascal Hitzler, Rudi Studer and then they are all different individuals or in Manchester syntax and in turtle syntax, sorry, we can say we can define a set of things using the square brackets. 中文初译: 在曼彻斯特语法中,可以这么说。你可以说不同的个人, Jorg Zucher、Pascal Hitzler、Rudi Studer, 然后他们都是不同的人, 或者在曼彻斯特语法和海龟语法中,抱歉,我们可以说我们可以用方括号来定义一组东西。 00:45:06.770 EN: And then we have the distinct members and then a list of things here, Pascal Hitzler, Rudi Studer, Jorg Zuhrer. 中文初译: 然后我们有不同的成员 和一份清单 这里的东西,帕斯卡尔・希茨勒,鲁迪・斯图德, 约格・祖尔赫。 00:45:13.930 EN: So they are all different individuals then. So there is a small task here. Is there a contradiction? So let's maybe together look at this example. 中文初译: 因此,他们都是不同的个人。所以这里的任务很小。有矛盾吗?让我们一起来看看这个例子。 00:45:33.020 EN: We have a class organization. We have a class office. Office is a subclass of WGS 84 spatial things. So spatial, this WGS 84 is an ontology for describing spatial things and there is the class defined spatial thing in there. 中文初译: 我们有一个阶级组织。我们有一个阶级办公室。办公室是WGS 84空间事物的子类。所以空间,这个WGS 84是一个用于描述空间事物的本体论, 那里有定义的空间事物。 00:45:53.900 EN: So office is a spatial thing. There is an object property has office and an organization can have an office. That's why the domain of has office is organization and the range office. 中文初译: 办公室是一个空间问题。有一个物体财产拥有办公室,一个组织可以拥有办公室。这就是为什么拥有办公室的领域是组织和范围办公室的原因。 00:46:10.520 EN: Then we have imported axioms for longitude and latitude. So WGS 84 latitude has the domain spatial thing and the same with longitude. 中文初译: 然后我们进口了经度和纬度的轴。所以WGS 84 纬度有空间域,经度也一样。 00:46:26.000 EN: Yeah, a spatial thing should be locatable on the map and that's why we have longitude and latitude for identifying something on the map. 中文初译: 是啊,一个空间的东西 应该可以定位在地图上 这就是为什么我们有经度和纬度 来识别地图上的东西 00:46:40.030 EN: And then we define an individual company xy is of type organization. It has office xy and then we have a comma. 中文初译: 然后我们定义一个单独的公司Xy是类型的组织。它有办公室xy,然后我们有一个逗号。 00:47:00.780 EN: That means we have further facts about the company here, the latitude and the longitude. So this would actually not be correct because the longitude and the latitude should not be applied to the organization but it should be applied to the office. 中文初译: 这意味着我们掌握关于公司、纬度和经度的更多事实。因此,这实际上是错误的,因为经度和经度不应适用于组织,而应适用于办事处。 00:47:26.520 EN: Yeah, because otherwise if we apply it to the organization, organization, since we have defined the domain of spatial thing, the company xy would become also a spatial thing. 中文初译: 是的,因为不然的话,如果我们把它应用到组织,组织, 因为我们定义了空间事物的领域, 公司xy也会成为空间事物。 00:47:43.340 EN: So this is quite important. And companies are usually not spatial things. They can have different offices even in different countries on different continents. 中文初译: 因此,这是非常重要的。公司通常不是空间事物。它们可以拥有不同的办公室,即使是在不同大陆的不同国家。 00:47:59.120 EN: So there you see that it's important to model things right and to be careful what you apply properties on. So then we have nominals or enumerations. 中文初译: 所以,你可以看到,重要的是要正确建模和小心使用属性。然后我们有标称或计数。 00:48:14.160 EN: So for example, we can say, we can define a class moon of Mars and we define this to be equivalent to a set of instances, Phobos and Deimos, which are the moons of Mars. 中文初译: 例如,我们可以说, 我们可以定义火星的一等月亮, 我们把它定义为 相当于一系列的例子, 即Phobos和Deimos, 它们是火星的月亮。 00:48:32.180 EN: And in Manchester syntax, we would write it as follows. So we write moons of Mars as a class and then semicolon that means moons of Mars is equivalent to, and then we have something in curly brackets that means this is basically a blank node, it's equivalent to a blank node. 中文初译: 在曼彻斯特语法中,我们会写如下。所以我们把火星的月亮写成一个类,然后是分号, 意思是火星的月亮就是等同的, 然后在卷轴括号里有东西表示这是一个空节点, 它相当于一个空白节点。 00:48:54.940 EN: This blank node is also of type class and it's one of Phobos and Deimos. So this is a list of things. So we define the class moons of Mars to be equivalent to an unnamed class and this unnamed class is constructed out of the set of these two individuals which are named here, Phobos and Deimos. 中文初译: 这个空白节点也是类型级的, 它是 Phobos 和 Deimos 之一。 因此这是一个列表。 所以我们定义了火星等级月亮相当于一个未命名的类别, 而这个未命名的类别是由这两类人组成的, Phobos 和 Deimos 。 00:49:28.930 EN: So here you see that the Manchester syntax is much more compact than the turtle syntax and even in the turtle syntax, we already simplified that a lot here. 中文初译: 所以这里,你可以看到曼彻斯特语法比海龟语法要紧得多。 即使在海龟语法中,我们已经简化了很多。 00:49:41.790 EN: So if we would write out all the triples, it would become even more verbose. So we can apply logical class constructors to build new classes and you know this logical conjunction, disjunction, negation and we can apply them to sets and that's why we can also apply them to classes. 中文初译: 所以如果我们写出所有的三重, 它就会变得更变的多动。 所以我们可以应用逻辑级构建器来建造新的班级, 你知道这种逻辑连结, 脱节,否定, 我们可以把它们应用到设置上, 这就是为什么我们也可以把它们应用到课堂上。 00:50:11.130 EN: So the conjunction is intersection of, so we can build the intersection of two classes. Then we have the disjunction, which means a union. 中文初译: 所以连接是交叉的, 所以我们可以建立两个等级的交叉。 然后我们有了脱节, 这意味着一个结合。 00:50:25.230 EN: We can build a union of two classes. There is negation. We can build complement of a class. So using these construct class constructors, we can construct more complex classes from simpler classes. 中文初译: 我们可以建立一个由两个班组成的联盟。 存在否定。 我们可以构建一个班级的补充。 因此,利用这些建筑类的建筑师, 我们可以从更简单的班级中建造更复杂的班级。 00:50:45.470 EN: So let's look at some example, for example, intersection. So let's maybe look at the turtle syntax here. We have moon of Mars is equivalent to the class and then we define a class here using a blank node. 中文初译: 让我们来看看一些例子,例如交叉点。让我们来看看这里的海龟语法。 我们有火星月亮相当于这个班级, 然后用空白节点来定义一个班级。 00:51:04.430 EN: So in this curly brackets, these are always blank nodes, which is also a class and then an intersection of moon and object near Mars. 中文初译: 所以在这个卷轴括号里, 这些总是空白的节点, 这也是一个等级, 然后是火星附近的月球和物体的交叉点。 00:51:18.550 EN: So an intersection of a list of things. So in round brackets, we have lists of things here and the intersection of moon and object near Mars is then equivalent to the class moon of Mars. 中文初译: 因此,一个事物列表的交叉点。在圆括号中,我们这里有各种事物的清单, 火星附近的月球和物体的交叉点相当于火星级的月球。 00:51:36.570 EN: So similarly, we can build union. Let's look again here at the Manchester syntax. We have boat is a subclass of, and then we have defined an anonymous class, which is of type class and then this anonymous class is a union of sailboat and motorboat. 中文初译: 类似地,我们可以建立工会。让我们再看看曼彻斯特的语法。我们有船是一个小类, 然后我们定义了一个匿名阶级, 属于型类, 而这个匿名阶级是帆船和机动艇的联盟。 00:51:57.070 EN: So we basically built a union of these two classes and define boat to be equivalent to this union or a subclass of this actually here, not equivalent. 中文初译: 因此,我们基本上建立了一个由这两个等级组成的联盟, 并定义了船相当于这个联盟或一个子类 它实际上在这里,不相等。 00:52:14.320 EN: So then there is also disjoint union so here is an RDF XML example. We have the class professor, it's a subclass of and then we define a new anonymous class to be a union of and then we have active and retired. 中文初译: 因此,也存在脱节的结合,所以这里有一个“RDF”“XML”的例子。我们有班级教授,一个小类,然后我们定义一个新的匿名班, 是一个工会, 然后我们有活跃和退休。 00:52:39.730 EN: So in Manchester syntax, it would look like that. So professor is a subclass of and we have this anonymous class, which is identified using a blank node and is a disjoint union of active and retired. 中文初译: 所以在曼彻斯特语法中,它会是这样的。所以教授是一个子类,我们有一个匿名班级, 它用空白节点来识别, 是一个活跃和退休的脱节。 00:52:59.190 EN: So in a simple union, an instance could belong to several of the components of the classes in a disjoint union instances should only belong to one of them. 中文初译: 所以在一个简单的结合中, 一个例子可能属于 脱离合并情况下 班级的几个组成部分, 应该只属于其中之一。 00:53:17.110 EN: So this is a bit syntactic sugar, you could also define it as a union and then define the elements of the union to be disjoint from each other. 中文初译: 所以这是一个有点合成的糖, 你可以把它定义为一个工会, 然后定义联盟中互相脱节的要素。 00:53:28.750 EN: So this makes it a bit more compact. And then finally complement, let's maybe look right away at the turtle syntax. 中文初译: 这样它就会更紧凑一点。最后,我们来看看海龟语法 00:53:38.090 EN: So we have faculty staff is an all class and it's a subclass of and then we have defined again an anonymous class, RDF type class and this is a complement of publication. 中文初译: 因此,我们有教职员 是一个班级,它是一个小班, 然后我们又定义了 一个匿名班,“RDF”类, 这是出版物的补充。 00:54:00.460 EN: So everything which is not a publication is this anonymous class and then we define faculty staff to be a subclass of that. 中文初译: 所以所有不是出版物的东西 都属于匿名课程 然后我们把教职员工定义为其中的一个子类 00:54:11.540 EN: So let's look at some modeling example. So if we want to model that all persons are either male or female, we could do this as follows. So we write male subclass of person, female subclass of person, female is disjoint with male and then we can define person to be equivalent to an anonymous class, 中文初译: 所以让我们来看看一些示范性的例子。所以如果我们想模拟所有的人都是男的或女的, 我们可以做如下的事情。 所以我们写出男性子类的人,女性子类人,女性子类人, 女性与男性脱节, 然后我们可以定义一个人相当于一个匿名类人, 00:54:42.970 EN: which is the union of male and female. Why do we need the second part here? Yeah, because the first part defines only that male and females are disjoint. 中文初译: 我们为什么在这里需要第二部分?是的,因为第一部分只界定了男女脱节。 00:54:57.270 EN: But the first part would not say there might be a third sex and the third sex might also be a subclass of person. 中文初译: 但第一部分不会说可能有第三性,而第三性也可能是一个小类的人。 00:55:07.190 EN: So the first part of this definition doesn't exclude that there are some third, that there are some other types of people who could also be persons. 中文初译: 所以这个定义的第一部分 不排除有三分之一, 还有一些其他类型的人 也可以是人。 00:55:21.930 EN: Only with the second part, we actually close the set of persons that it's a union of males and females, yeah. After we added this part, there can be no other instances of person which are not males or females. 中文初译: 只有第二部分,我们才真正关闭了这组人, 即男女的联盟, 是的。在我们添加这部分之后, 没有其他人不是男性或女性。 00:55:43.330 EN: That's why the second part is very important. Although it's maybe not completely correct anymore, it's a bit unclear. 中文初译: 这就是为什么第二部分非常重要。虽然它可能不再完全正确,但有点不清楚。 00:55:52.270 EN: So this binary definition of sexes, I think is not valid anymore. So in that case, you would remove the second part here. 中文初译: 因此,我认为对性别的二元定义不再有效。在这种情况下,你将在这里删除第二部分。 00:56:00.930 EN: And if you have this first part, then there could be also a third sex. Then we have role limitations, like all values from. 中文初译: 如果你有这第一部分,那么也可能有第三性。然后我们的角色限制,就像所有价值观一样。 00:56:13.510 EN: So we can define classes by roles. And here we see an RDF XML example. So we define a class exam. It's a subclass of something. 中文初译: 这样我们可以按角色来定义等级。 我们在这里看到一个“ RDF” “XML ” 示例。 所以我们定义了一个类考试。 这是一个子类, 它是一个小类。 00:56:23.970 EN: And the something is a restriction on the property examiner. And all values of examiner must be professor. And then the restriction ends. 中文初译: 检查者的所有价值都必须是教授。然后限制就结束了 00:56:34.550 EN: So this basically selects all instances which have the property examiner and where all values of this property examiner belong to the class professor. 中文初译: 因此,这基本上选择了拥有财产检验员的所有实例, 以及该财产检验员的所有价值都属于班级教授的事例。 00:56:50.330 EN: And the set of these instances then is the set of this restrictions here, restriction and we define exam to be a subclass of it. 中文初译: 然后,这些例子的一组情况 就是这些限制, 限制,我们把考试定义为它的一个子类。 00:57:02.510 EN: So here we have it also in turtle syntax. So we have exam subclass of an anonymous class, basically a blank node. 中文初译: 我们这里也有海龟语法。所以我们有一个无名类的考试小类, 基本上是一个空白节点。 00:57:13.910 EN: And this anonymous class is an old restriction class, a restriction on the property examiner, and all values of examiner must belong to the class professor. 中文初译: 这个匿名阶级是旧的禁制阶级, 对财产审查者的一种限制, 考试者的所有价值必须属于教授。 00:57:26.910 EN: So you can imagine this a bit like query. You have a knowledge base of professors, of exams, of examiners, of students, and then you select all instances which have the property examiner and where the values, all values of this property belong to the class professor. 中文初译: 所以,你可以想象这有点像查询。你有一个教授、考试、考试、学生的知识库, 然后你选择了拥有财产审查员的所有实例, 以及这些财产的价值和所有价值都属于班级教授的事例。 00:57:54.950 EN: And this set of instances is then what we have defined here. And then you define exam to be a subclass of that. 中文初译: 这组例子就是我们在这里定义的。然后你将考试定义为其中的一个子类。 00:58:04.790 EN: You basically construct a class from selecting instances by the properties and the property values. And this also says that all examiners must be professors. 中文初译: 基本上您从选择属性和属性值来构建一个类。 这还表明所有考试者都必须是教授。 00:58:18.810 EN: So cardinality, there's also cardinality restrictions. So let's look at a turtle example. We have exam is in our class. 中文初译: 因此,最基本性,也存在最基本的限制。让我们来举个海龟的例子。我们有考试是班级的。 00:58:30.180 EN: It's a subclass of, and then we have again an anonymous class definition here. It's an, and this is a restriction on the property examiner. 中文初译: 这是一个子类,然后我们又有一个 匿名的分类定义。这是一个,这是对财产审查的限制。 00:58:43.490 EN: And then we have defined max cardinality too. So we basically select all instances which have the property examiner which is used maximum two times. And these are then the exams, the set of exams or exams must be a subset of that. 中文初译: 然后,我们定义了最大的最基本性。所以我们基本上选择了所有有财产检验员的事例,他们最多使用两次。然后这些是考试,考试或考试必须包含其中的一个子集。 00:59:18.080 EN: So it, so this defines basically that each exam can have at most two examiners. And there's also min cardinality or exactly. 中文初译: 因此,这基本定义了每个考试最多可以有两个考官。 并且还有次要的或确切的特征。 00:59:32.560 EN: So here we showed you an example of max cardinality but you can also replace that with min cardinality or with exactly. 中文初译: 因此我们在这里给你们展示了一个最基本性的例子, 但你也可以用最基本或确切的表示来取代。 00:59:44.620 EN: So a performance, let's have a look at another example. A performance requirement of a software product is a requirement which is created by customers. It leads to a system requirement. 中文初译: 因此,一个性能,让我们来看看另一个例子。软件产品的性能要求是客户的要求。它导致系统要求。 00:59:58.740 EN: So this would be something from. Software engineering, maybe some of you have attended the software engineering lecture. 中文初译: 所以这是来自软件工程,也许你们中有些人参加了 软件工程讲座。 01:00:05.380 EN: And there we have requirements engineering, and we want to model this in an ontology. So let's, here we have only the Manchester syntax example. 中文初译: 我们有需要的工程学,我们想用本体论来做这个模型。所以,让我们在这里,我们只有曼彻斯特语法的例子。 01:00:18.300 EN: We have performance requirement, a subclass of requirement, and created by only customer leads to some system requirement. 中文初译: 我们有业绩要求,一个次级要求类别, 由只有客户才创造的, 导致某种系统要求。 01:00:32.000 EN: So that's how we could model this in using Manchester syntax. Then we have another restriction, role limitations. 中文初译: 这就是我们用曼彻斯特语法来模拟这个模式的方法。然后我们又有了另一个限制,角色限制。 01:00:48.970 EN: And for example, has value. So let's look at this example we define in our class here. This is an RDF XML example. Exam at Fenrich, so by Professor Fenrich, is equivalent to a class which we define here. 中文初译: 例如,有价值。让我们看看我们在这里的类中定义的这个例子。 这是一个“RDF” XML 的例子。 Fenrich的考试,由Fenrich教授进行,相当于我们在这里定义的一个类别。 01:01:07.890 EN: And this class is a restriction on the property examiner has value Fenrich. So what does this do? This restriction basically selects all the instances which have the property examiner and where one of the values, not we had all values before, now just one of the values must be Fenrich. 中文初译: 这个类别是对财产检验员的限制, 其价值是 Fenrich 。 因此它能做什么呢? 这个限制基本上选择了所有有财产检验员的事例, 其中有一个价值,而不是我们以前拥有的所有价值, 现在只有其中之一必须是 Fenrich 。 01:01:34.730 EN: Because exams can have several examiners, so we select those where at least one is of the examiners is Fenrich. 中文初译: 因为考试可以有数名考官, 所以我们选择其中至少有一名考官是芬里希。 01:01:44.690 EN: And define this to be the class exams at Fenrich. In Manchester syntax, sorry, in Turtle syntax here below, it would look like this. 中文初译: 并定义这是Fenrich的班级考试。在曼彻斯特语法中,抱歉,在下面的“Turtle”语法中,它看起来是这样的。 01:01:56.490 EN: Exams, actually it should be exam at Fenrich, RDFS equivalent class. And then we define it anonymously here. This is a restriction on the property examiner. 中文初译: 考试,实际上应该是Fenrich的考试,“RDFS”等同等级。然后我们在这里匿名定义它。这是对财产审查员的限制。 01:02:11.990 EN: So we select all instances which have the property examiner and have the value Fenrich. And these instances then create a new anonymous class, this one here, this restriction class. 中文初译: 所以,我们选择所有有财产检验员和Fenrich价值的示例。然后这些实例创建了一个新的匿名类, 这个是这个限制类。 01:02:26.010 EN: And we say that exam at Fenrich is equivalent to this anonymous class. And then we have domain and range. Maybe we don't need to spend much time on that because it's exactly the same as in RDF schema. 中文初译: 我们说在芬里希的考试相当于这个匿名班。 然后我们有域和范围。也许我们不需要花很多时间去参加, 因为它与“RDF”计划完全相同。 01:02:46.210 EN: So we can define an object property belongs to and the range organization. So here we have belongs to type object property, then the range organization. So it looks in Turtle syntax. 中文初译: 所以我们可以定义一个属于对象属性和范围组织。 所以这里我们属于输入对象属性, 然后是区域组织。 它在“ Turtle” 语法中查看 。 01:03:04.210 EN: And we could also write this now more complex as a owl restriction. So we write owl thing is a subclass of a restriction on property belongs to all values from organizations. 中文初译: 并且我们也可以写这个现在更复杂的 猫头鹰限制。所以我们写猫头鹰是限制财产的子类 属于组织的所有价值。 01:03:30.540 EN: So what does it mean? It basically means that if we select all instances which have the property belongs to, then all the values of these properties must be organizations. 中文初译: 那么这意味着什么?它基本上意味着如果我们选择拥有财产的所有实例,那么这些财产的所有价值必须是组织。 01:03:50.690 EN: And since we define owl thing as a subclass of that, it means there can be nothing else because owl thing is already the set of all instances. 中文初译: 既然我们把猫头鹰 定义为一个小分类, 这意味着没有别的东西了 因为猫头鹰已经是所有例子的组合。 01:04:01.970 EN: So how do you model the domain D of a property p without using the domain keyword? And as a hint, whenever p occurs at least once on something, this thing is of type D. For at least once, think of cardinality constraints or of existential quantification. 中文初译: 那么您如何在不使用域关键字的情况下模拟属性 p 的域 D ? 作为提示, 当 p 发生至少一次时, 它就属于 D 类 。 至少一次, 想想基本限制或存在量化 。 01:04:24.570 EN: So we could use this mean cardinality constraint to represent also the domain. So what do we have here? Domain and range. 中文初译: 因此,我们可以使用这个隐含的基点限制来同时代表域。 那么我们这里有什么? 域和范围。 01:04:38.190 EN: So we have defined an object property belongs to. The range is organization. Then we define number five. And we use belongs to primes. 中文初译: 所以我们定义了属于一个物体的属性。 范围是组织。 然后我们定义了第五个。 我们使用的是质数。 01:04:55.700 EN: Five is a prime number. So what would follow here? Since we have range organization, primes would become an organization. 中文初译: 五是质数。 那么接下来会是什么? 既然我们有范围组织, 质数将成为一个组织。 01:05:12.680 EN: So in RDF, it works. RDF schema owl, it works a bit differently than, for example, in a database. In a database, if you define a schema, everything must fit to the schema. 中文初译: 在“RDF ” 中,它有效。“RDF ” Schema 猫头鹰(RDF ) , 它的工作方式与数据库的工作方式略有不同。在数据库中,如果你定义一个方案,那么所有的东西都必须适合这个方案。 01:05:24.980 EN: Here in RDF, it's a bit differently. We define constraints, and then a reasoner will try to make the knowledge base fit to the constraints. 中文初译: 在“RDF”中,它有点不同。我们定义了限制,然后一位理智者将试图使知识库适应这些限制。 01:05:37.050 EN: If we define, for example, range of belongs to to be an organization, and then we use belongs to to indicate that five belongs to the set of prime numbers, then a reasoner would infer that primes is an organization. 中文初译: 例如,如果我们界定一个组织所属的范围,然后我们用它来表示五人属于质数组,那么一个理性者就会推断质数是一个组织。 01:05:56.260 EN: So again, owl and RDF schema, they restrict or they define axioms. But of course, you can still write all kinds of triples. 中文初译: 同样,猫头鹰和RDF的图案 限制或定义了正弦。当然,你仍然可以写出各种三重曲。 01:06:10.500 EN: But a reasoner would then come to awkward conclusions, sometimes even to inconsistencies in the knowledge base. 中文初译: 但理智者会得出尴尬的结论,有时甚至会发现知识库的不一致。 01:06:19.500 EN: This here would not yet be an inconsistency. If you add another axiom that primes, for example, the set of prime numbers is disjoint from organization, you could add this. 中文初译: 这里不会出现不一致。 如果您添加另一个基本值, 例如, 质数组与组织脱节, 您可以添加此点 。 01:06:31.660 EN: Then this would lead to an inconsistency. So what's important to remember is that RDF, RDF schema owl, they don't really prohibit things. Domain and range doesn't prohibit that you can use other resources as domain and range. 中文初译: 这会导致不一致。所以重要的是记住,“RDF ” 、 “RDF ” 的猫头鹰,它们并不真正禁止事物。域和范围并不禁止您使用其他资源作为域和范围。 01:06:51.820 EN: But the result of that would be that they become part of the domain and range classes. And this could lead to inconsistencies. And that helps you. 中文初译: 但结果将是它们成为域和范围类别的一部分。这可能导致不一致。这对你有帮助。 01:07:03.520 EN: If a reasoner detects such inconsistencies, then you can do some back fixing on your ontology and you can identify where the error is. 中文初译: 如果一个理智者发现有这种不一致之处,那么您就可以对本体论做一些后退修补,并可以识别出错地点。 01:07:13.620 EN: Here in that case, you should not use belongs to, but you should use another property, define different properties. 中文初译: 在此情况下, 您不应该使用属于您的, 但是您应该使用其它属性, 定义不同的属性 。 01:07:26.880 EN: Here we have the example again in turtle. So things are not prohibited, but they might lead to unintended consequences. 中文初译: 这里我们再次以海龟为例。所以事情并不被禁止,但可能导致意外后果。 01:07:38.100 EN: That's maybe very important to remember. All RDF schema do not prohibit things, but might lead to unintended consequences. 中文初译: 这也许非常重要,值得记住。所有RDF的计划并不禁止事物,但可能导致意外后果。 01:07:48.720 EN: And the reasoner might then hint you and give you some hints of what the reason is, and then you can fix your ontology. 中文初译: 然后理智者可能会暗示你 给您一些原因的提示 然后你就可以修复你的本体论了 01:08:00.640 EN: So role properties, maybe we skip this RDF XML example. So basically here we define hasColleague is a transitive and symmetric property. 中文初译: 所以角色属性, 也许我们跳过这个“ RDF” “XML ” 示例。 因此基本上这里我们定义的“ 共产主义” 是一个过渡性和对称性属性。 01:08:18.500 EN: And then we define projectManager is a functional property. There is only one manager for a project. isProjectManager4 is an inverse functional property. 中文初译: 然后我们定义项目管理员是功能性财产。 只有一个项目管理员。 是工程管理员4 是反函数性财产。 01:08:33.200 EN: And it could be also defined as an inverse of hasProjectManager, right? And then we have here zur and auer, hasColleague, Sebastian Tramp, and Jens Lehmann isProjectManager for triplify. 中文初译: 也可以被定义为“项目经理”的反面,对吗?我们这里有“zur”和“auer ” , “Colgloble,Sebastian Tramp 和“Jens Lehmann”是“三重组合”的项目经理。 01:08:47.380 EN: That was a project we were doing. And then we have another project. Onto wiki, you can infer, ah, actually there was zur and auer and auer zur, and let's look at that. Where did we find this here? 中文初译: 那是我们正在做的一个项目,然后我们又做了另一个项目。关于维基,你可以推断, 啊,实际上那里有zur和auer zur和auer zur,让我们看看。我们在哪里找到这个的? 01:09:02.240 EN: Here we have the project onto wiki, as projectManager, and then we have zur and auer, and then we have another statement, hasProjectManager, auer zur. 中文初译: 这里我们有Wiki的项目,作为项目经理, 然后我们有zur和auer, 然后我们还有另一个声明, ProjectManager, auer zur。 01:09:15.360 EN: Since we have defined hasProjectManager as a functional property, these two project managers here must be the same. Refer to the same person. 中文初译: 由于我们把项目管理员界定为一种功能性财产,这两个项目经理必须是同一人。 01:09:28.220 EN: So functional property does not prohibit that you use the property several times, but the result of that is that the values of this property, the objects, must refer to the same person. 中文初译: 因此,功能性财产并不禁止您多次使用该财产,但结果就是该财产的价值,即物品,必须指同一人。 01:09:42.440 EN: And that a reasoner might then conclude and infer, and then you can check if it was intended or unintended. So in OWL2, there are only, there are also negative facts. 中文初译: 然后一个理性者可以得出结论和推断, 然后你就可以检查这是有意的还是无意的。所以在“OWL2 ” 中,只有,还有负面的事实。 01:10:00.300 EN: So in Manchester syntax, you can define an individual bill and then the fact not as wife Mary. In Turtle syntax, this would look as follows. 中文初译: 所以在曼彻斯特语法中,你可以定义一个单项法案, 然后事实就不再是玛丽的妻子。在“Turtle”语法中,这看起来是这样的。 01:10:13.260 EN: We have a blank node here. Type OWL negative property assertion. Then we have the source individual bill. The assertion property has wife and the target individual Mary. 中文初译: 我们这里有一个空白节点, 输入“ OWL” 否定的财产主张。 然后我们拿到了源个人帐单。 声称的财产有妻子和目标个人Mary。 01:10:30.420 EN: So we basically say that bill has wife Mary is a negative property assertion. So basically that this property does not hold. 中文初译: 因此,我们基本上说法案有妻子玛丽是一个否定的财产主张。所以基本上,这些财产是不能持有的。 01:10:41.300 EN: So it's a bit complicated. You need four triples to express that one fact doesn't hold. So here we have a modeling example. 中文初译: 所以这有点复杂。你需要四个三倍来表达一个事实是站不住脚的。所以我们这里有一个示范的例子。 01:10:50.860 EN: There are male and female persons and people have names. So we have male person, a subclass of person, female person, subclass of person, has name as a data type property, has the domain person and the range string. 中文初译: 男女都有姓名,所以我们有男性、一个小类人、女性、小类人、以数据类型财产命名、拥有域名和范围字符串。 01:11:09.040 EN: Then we define John as a male person and can have, he has the name John. We add it as a string here, indicated by the data type, XSD data type. 中文初译: 然后我们把约翰定义为男性,可以拥有,他有约翰的名字。我们在此添加一个字符串,用数据类型XSD数据类型来表示。 01:11:24.300 EN: And we have Jill is a female person and her name is Jill. So this is basically an overview of OWL one. OWL became a W3C recommendation in 2004. OWL two, then in 2009. 中文初译: 我们有吉尔是个女性,她的名字叫吉尔。这基本上是对“OWL”的概述。“OWL”在2004年成为W3C的建议。“OWL”二,然后在2009年。 01:11:46.000 EN: OWL two is fully backward compatible with OWL one, but there are different subsets and some syntactic sugar was added in OWL two. 中文初译: “OWL”二号与“OWL”一号完全反向兼容,但有不同的子集,在“OWL”二号中增加了一些合成糖。 01:12:01.030 EN: And in OWL two, there was also a bit more expressivity like keys, property chains, richer data types, data ranges, qualified cardinality restrictions, asymmetric reflexive and disjoint properties and further enhanced annotation capabilities in OWL two. 中文初译: 在“OWL”第二版中, 也有比钥匙、财产链、更丰富的数据类型、数据范围、限定的基点限制、不对称反射和脱节性能以及进一步增强“OWL”第二版中的注解能力等更多一些表达性。 01:12:24.310 EN: So, and then there are some subset defines. Basically, OWL full allows unlimited use of all OWL and RDF language elements. 中文初译: 所以,然后有一些子集定义。基本上,“OWL” 完全允许无限制地使用所有“OWL”和“RDF”语言元素。 01:12:40.080 EN: So for example, you don't have to separate classes, roles and individuals. So that's why OWL thing is the same as RDFS resource. 中文初译: 比如说,你不必区分班级、角色和个人。这就是为什么“OWL”和“RDFS”资源是一样的。 01:12:55.190 EN: OWL class is the same as RDFS class. OWL data type property is a subclass of all object property and object property is the same as RDF property. 中文初译: “OWL”类与“RDFS”类相同。“OWL”数据类型属性是所有物体财产的子类,而物体财产与“RDF”类相同。 01:13:08.290 EN: And then there are some profiles in OWL two, which were added OWL QL, EL and RL. And the idea is to restrict the use of OWL that you are not allowed to use all the language features so that the reasoning and the inference could be performed faster. 中文初译: 然后在“OWL” 2中有一些描述, 添加了“OWL” QL、EL和RL。 目的是限制使用“ OWL” , 即不允许您使用所有语言特征, 这样推理和推论可以更快地进行 。 01:13:30.030 EN: So the design principle was to identify sub languages of OWL two that are still implementable in polynomial time where the inferencing can be done in polynomial time. 中文初译: 因此设计原则是确定“OWL”二类的子语言,在多元时间仍然可以执行,在多元时间可以作出推论。 01:13:44.410 EN: So, and in general, this allows negation and disjunction because that complicates reasoning and is rarely needed. 中文初译: 因此,一般来说,这允许否定和脱节,因为这使推理复杂化,而且很少需要。 01:13:53.810 EN: Negation and disjunction are difficult to handle. You can imagine if you combine all these features which we have learned today, this inferencing can become quite complex and cumbersome. 中文初译: 忽略和脱节是很难处理的。 你可以想象,如果你把我们今天学到的所有这些特征结合起来,这种推论可能会变得相当复杂和繁琐。 01:14:08.000 EN: So OWL QL is query, QL stands for query language. And it's designed for query answering using SQL reviting on top of relational databases. 中文初译: 所以“ OWL” QL 是查询, QL 代表查询语言。 它设计用于使用“ SQL” 在关系数据库上方进行查询回答 。 01:14:26.100 EN: Subclasses can only be class names or existentials with unrestricted filler. Superclasses can be class names, existentials or conjunctions with superclass filler or negations. 中文初译: 子类只能是无限制填充者所属的类名或存在者。 超级类可以是类名、存在者或与超级填充者或否定者相关联者。 01:14:40.220 EN: So that's a restriction basically of OWL. And here you see what is allowed, for example, subclass relationship or this restriction here is allowed, but some other things, this is description logic syntax. 中文初译: 这就是“OWL”的限制。在这里,你可以看到什么是允许的, 比如子类关系或者这个限制在这里是允许的, 但是其他一些东西,这是描述逻辑语法。 01:14:56.460 EN: I have not introduced that yet, but some definitions, more complex definitions are not allowed in QL. 中文初译: 我尚未介绍这一点,但一些定义、更复杂的定义在质量L中是不允许的。