Ontology is an electronic engineering term.staycomputer scienceAndinformation scienceIn theory, ontology refers to a "formal, clear and detailed description of the shared concept system".Ontology provides a shared vocabulary, that is, those existing in a specific domainobjecttypeOr concept and its attributes andinterrelation；In other words, ontology is a special typeGlossary, with structural characteristics, and more suitable forcomputer systemIntermediate use;In other words, ontology is actually about specificfieldOne setconceptAnd between themrelationshipFormal representation of.Noumenon is that people take their own interestsfieldOfknowledgeIt is written for material using the ontology principle of information scienceworks。Ontology can generally be used to implementreasoning, can also be used to define the domain (that is, to model the domain).In addition, sometimes people also refer to "ontology" as "ontology".[4]
English terminology“ontology”It originates from the field of philosophy and has been used in many different ways.In the field of computer science, its core meaning is a model used to describe a set of object types（conceptOr sayclass）, attributes, and relationship types.Although different ontologies have different exact terms for these components, they are all indispensable basic elements of an ontology.Generally speaking, people generally expect that the features of the model in the ontology should be very similar to the corresponding real world.
As far as computer science and philosophy are concerned, the common ground between the two ontologies is that they are expressed according to a certain category systementity、concept, events and their attributes and interrelationships.In these two fields, there are various problems concerning ontology relativity (for example, in the field of philosophyQuineandSaul Kripke , in the field of computer scienceSova andGaleno）, people have done and are doing a lot of work;Moreover, people are also discussing whether the standardized ontology has vitality (for example, the discussion on fundamentalism in the field of philosophy, and the discussion onCycProject discussion).The differences between the two are, to a large extent, just a matter of emphasis.Compared with researchers in the field of computer science, philosophers are less concerned with creating fixedcontrolled vocabulary 。However, computer scientists are less involved in discussions about the first principles (for example, whether there areInvariable essence(fixed essences), or the ontology of entities must be more basic than that of processes).[1]
history
Announce
edit
Historically, ontology originated fromPhilosophyIs called“metaphysics”Branch of.Metaphysics is concerned withrealityOfessence, that isexistenceThe essence of.As a basic branch of philosophy, metaphysics focuses on the analysis of various types or modes of existence, especiallyCophaseAndParticular The relationship betweenIntrinsic and extrinsic propertiesThe relationship betweenessenceAndexistenceRelationship between.The traditional goal of ontological exploration activities, in particular, is to divide the real world into natural categories of objects at key points in order to reveal those basic categories or categories.
In the second half of the 20th century,philosopherThey had extensive discussions on the feasible methods or approaches of how to build ontology, but in fact they did not really "build" any very detailed ontology.In contrast,Computer scientistAt that time, we were building some large and stable ontologies (such asWordNetandCyc）However, relatively speaking, there is little debate on how to build ontology.
Since the mid-1970s,artificial intelligence(short forAI）Researchers in the field have realized that knowledge acquisition is the key to building a powerful AI system.AI researchers believe that they can create new ontologies intocomputational modelTo achieve specific types ofAutomatic reasoning。In the 1980s, AI began to use terminologyontologyTo simultaneously refer to a theory about the modeled world and a component of the knowledge system.With the inspiration from philosophical ontology, some researchers then regard computer ontology as a kind ofApplied philosophy。
In the early 1990s,Tom Gruber Published a widely quoted web page andpaper《Design principles for knowledge sharing ontology》。The reason why this paper is so popular is that Grubercomputer scienceterm“ontology”A prudent definition of.Gruber uses this term to refer to aConceptual system（Conceptual expressionorConceptualization process）（conceptualization）Ofdetailed description（specification，Explicit statement）。In other words, just like the formal specification (formal specification) of a specific program, an ontology is a description of the concepts and relationships that may exist relative to an agent or a group of agents.This definition is consistent with the use of "ontology" as a "set of concept definitions", but it is relatively more general.However, the meaning of the word here is different from that of philosophyusage。
noumenonIt is often equivalent to those whodefinitionAnd the hierarchical structure of taxonomy formed by the classification relationship, but the ontology is not necessarily limited to this form.At the same time, ontology is not limited to conservative definitions (that is, traditionallogicThe definitions in the sense that they introduce and adopt aretermWithout adding any knowledge about the real world).When we want to specify a concept to be expressed clearly and in detail, we need to declare a number ofaxiomSo as to restrict and limit the possible interpretations of the defined terms.
In the first few years of the 21st century,cognitive scienceOfInterdisciplinaryThe project has been bringing scholars in these two fields closer together.For example, someone talked about "the computerization transformation in the field of philosophy".In this transformationphilosopherWe analyze the formal ontology compiled in the field of computer science (sometimes even directly used in workComputer software）At the same time, researchers in the field of computer science are also increasingly citing the research work of philosophers who are committed to ontology (sometimes even the direct result of the latter method).However, there are still many scholars who are not compliant in these two fieldscognitive scienceThis trend of development, and still continue to carry out their own work independently of each other, respectively engaged in their own concerns.[1]
Elements of ontology
Announce
edit
As far as the existing ontologies are concerned, no matter what language they are expressed in, they have many similarities in structure.As mentioned earlier, most ontologies describe individuals (instances), classes (concepts), attributes, and relationships.In this section, we will discuss these elements of ontology in turn.
Common ontology elements include:
Individual (instance): basic or "underlying" object.
Class:aggregate(sets), concept, object type or kind of things.
Attributes: attributes, characteristics, characteristics, characteristics and parameters that objects (and classes) may have.
relationship: The possible way of association between classes and individuals.
Function term: In a declaration statement, it can be used to replace the complex structure formed by the specific relationship of specific terms.
Constraints (restrictions): A formal description of the conditions that must be established to accept an assertion as input..
Rules: statements in the form of if then statements that describe logical inferences that can be derived from certain assertions in a specific form.
Axiom: Assertions (including rules) in a specific logical form together constitute the whole theory described by its ontology in the corresponding application fields.This definition is different from the "axiom" in production grammar and formal logic.In these disciplines, axioms only include those asserted astranscendentalStatement of knowledge.As far as the usage here is concerned, "axiom" also includes the theory derived from axiomatic statements.
event(philosophy) a change in attributes or relationships.[1]
Domain ontology
Announce
edit
Domain ontology(Domain ontology or domain specific ontology, that isDomain specific ontology）What is modeled is a specific domain, or part of the real world.Domain ontology expresses the special meanings of those terms that are suitable for the domain.For example, takeenglishword“card”Say.aboutpokerThe ontology of the domain may give the word“Play poker”Aboutcomputer hardwareThe domain ontology may be endowed with“Punch card”And“Video card”Means.
Upper body(Upper ontology or foundation ontology, that isBasic ontology）It refers to a model composed of common objects that are universally applicable in various domain ontologies.Included inCore vocabulary, which can be used to describe objects in a set of domains.There are several ready to use standardized upper ontologies, includingDublin Core、General formal ontology（General Formal Ontology，GFO）、OpenCyc/ResearchCyc、Recommend upper merge ontology(Suggested Upper Merged Ontology, SUMO) andDOLCE。In addition, some people believe thatWordNetIt belongs to the upper ontology, but actually it is not an ontology:WordNetJust oneclassification(taxonomy) and onecontrolled vocabulary The unique combination formed (see the above content on "attributes").
GellishOntology is an example of the combination of upper ontology and domain ontology.
Domain ontology uses a very special and often selective way to express concepts, so they often lack compatibility.With the expansion of systems that rely on domain ontology, they often need to combine different domain ontologies into a more general expression.For ontology designers, this poses a challenging problem.In the same field, due to cultural background, education level andideologyFor this fieldperceptionDifferent ontologies may appear due to different situations or different expression languages.
At present, the merging of ontologies that are not based on the same basic ontology is largely a manual process, which is time-consuming and costly.Those domain ontologies that use a set of basic elements provided by the same basic ontology to specify the meaning of domain ontology elements can realize automatic merging.There are many researches on general technical methods for ontology merging, but the research in this area is still at the theoretical level to a large extent.[1]
Body engineering, also known asOntology construction、Ontology preparationorOntology development, which aims to study the methods andMethodologyThe domain of.Ontology engineering includes ontology development process, ontology life cycle, ontology construction methods and methodologies, as well as toolkits and languages to support these aspects.
Ontology engineering aims to enable software applicationsenterprises and institutionsThe knowledge contained in the business operation process of the organization and specific fields becomes clear.The noumenon project is to solvesemanticsCaused by obstaclesInteroperabilityThe question points out a direction.For example, those related to businesstermAnd softwaredefinitionRelated barriers.Ontology engineering is actually a set of tasks related to ontology development and compilation in a specific field.[1]
The ontology generally adoptsOntological languageTo prepare.Ontological language, also known as“Ontological language”, which is used to compile ontologyFormal language。There are many such ontology languages, both proprietary and based onstandardOf:
Ordinary logic(Common logic)ISOStandard 24707;This is about aOntological languageThe family's technical specifications, in which the ontology languages can accurately transform each other.
CycThe project has its own, based onFirst-order predicate calculus, and has some higher-order extended ontology languages (i.eCycL）。
GellishThe language includes rules on its own expansion, so it integrates an ontology and an ontology language.
Rule exchange format(Rule Interchange Format, RIF) andF-Logic(F-Logic) can combine various ontologies and rules.
OWLWeb Ontology Language is used to write ontology statementslanguage。OWL's development inheritsRDFandRDFSAnd some early ontology language projects, includingOntology reasoning layer（Ontology Inference Layer，OIL）、DARPA Agent Markup Language(DARPA Agent Markup Language, DAML) andDAMLplusOIL。OWL is intended to be applied towebabove;Moreover, its constituent elements (classes, attributes and individuals) are defined as RDFresources, and adoptURIBe identified.[1]
Example of ontology
Announce
edit
Amino acid noumenon: a small noumenon about amino acids and their properties (in which concepts, attributes and relationships are provided with Chinese and English names, synonyms and abbreviations).
Basic Formal Ontology: a formal upper ontology designed to support scientific and technological research.
BioPAX: a department aimed at promoting data exchange andInteroperabilityThe body of.
Cell Cycle Ontology (CCO): an application ontology used to express cell cycle.
COSMO: It belongs to a basic ontology (the current version is OWL), which is designed to include all those that logically specify any fieldentityOfmeaningPrimitive concepts required at that time.It is intended to be a basic ontology, which can be used to complete other ontologies ordata baseThe transformation between different expressions in.At first, it was just a combination of the basic elements of OpenCyc and SUMO ontology.It has been supplemented by other ontology elements (types and relationships), so as to facilitate the inclusion of all the expressions of words in the Longman dictionary definition vocabulary.
General Upper Model: One part is used to connect the customer system with thenatural languageThe ontology with linguistic motivation that plays a mediating role between technologies.
Gellish English Dictionary: It is an ontology including dictionary and taxonomy;Among them, there is an upper ontology and a lower ontology, focusing on industrial and commercial applications in engineering, technology and procurement.See SourceForge'sOpen SourceProject.
GOLD:Language descriptionGeneral Ontology for Linguistic Description.
Open Biomedical Body Foundry(OBO Foundry): A set ofInteroperabilityReference ontology of.
Biomedical survey ontology(Ontology for Biomedical Investigations): used to describebiologyandclinicalAn open integrated ontology for research activities.
Plant Ontology: aboutBotanyStructure and growth/development stages.
The work of developing various ontologies for the Internet has bred those with search function to provide ontologiescatalog(directories) or listedservice。Such directories are called“Ontology library”。
The following are the static configurations of some manually selected ontologieslibrary：
CO-ODE project ontology library: provides some ontology examples and ontology resource links related to CO-ODE project.
The DAML Ontology Library stores historical ontology in DAML format.
The Protege Ontology Library includes a set of ontology in OWL format, framework based format and other formats.
SchemaWeb is a directory composed of RDF schemata expressed in RDFS, OWL and DAML+OIL formats.
The following are the directories andSearch Engines。These includeSearcher。
OntoSelect Ontology Library provides a variety of similar services for RDF/S, DAML and OWL ontologies.
Ontaria is "aSemantic WebData directory, focusing on RDF vocabulary and OWL ontology ".
SwoogleIt is a directory and search engine, applicable to all RDF resources available on the Internet, including various ontologies.[2]
Common ontology tools
Announce
edit
Ontology server
CO-ODE ontology browser: also known as "OWLDoc Server", namely, OWL ontology document server, used for dynamic generationHTMLOWLDoc document, similar toProtégéOWLDoc plug-in in ontology editor.
