1 / 68

PROTON PROTo ONtology: A Base Upper-Level Ontology For the Semantic Web Ivan Terziev Atanas Kiryakov SEKT Q4 Meeting

Ontotext Lab.

landry
Download Presentation

PROTON PROTo ONtology: A Base Upper-Level Ontology For the Semantic Web Ivan Terziev Atanas Kiryakov SEKT Q4 Meeting

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

    1. 16 February, 2012 PROTON PROTo ONtology: A Base Upper-Level Ontology For the Semantic Web Ivan Terziev Atanas Kiryakov SEKT Q4 Meeting Innsbruck, January 17-19, 2005

    2. Ontotext Lab #2/69 16 February, 2012 Contents How this presentation is organized: introduction, background info; a brief overview of PROTON architecture and modules; details of Top and some Upper module classes and properties; usage and extension guidelines future development issues; PROTON community process – ideas and organization; technical info – status, namespaces, web sites, etc. conclusions discussion Objectives: to shed more light on PROTON (current status); to add specific details over major classes and properties; to initiate a discussion on PROTON wrt a community process.

    3. Ontotext Lab #3/69 16 February, 2012 Introduction (I) Name. PROTON is an acronym for Proto Ontology ex-names: BULO (basic upper-level ontology), GO (generic ontology); not a Russian space rocket ? “proto” – used in the sense of “primary”, “beginning”, “giving rise to”, vs. “first in time” or “oldest”; connotations: positive, fundamental, elemental, “in favour of”, even romantic (like a science-fiction novel from the 60-ies) ? Intended usage. A Basic Upper-Level Ontology like PROTON - used for: ontology generation within SEKT WP1 (automatic extension of PROTON); metadata generation within SEKT WP2 (annotations and ontology population); knowledge modelling and integration strategy of a KM environment; generation of domain, application, and other ontologies.

    4. Ontotext Lab #4/69 16 February, 2012 Introduction (II) Documentation. SEKT deliverable D1.8.1: “Base-upper-level-ontology guidance” the first official version submitted M12 http://proton.semanticweb.org/D1_8_1.pdf Design principles: domain-independence; light-weight logical definitions; compliance with popular metadata standards; good coverage of concrete and/or named entities (i.e. people, organizations, numbers); no specific support for general concepts (such as “apple”, “love”, “walk”), however the design allows for such extensions

    5. Ontotext Lab #5/69 16 February, 2012 Some Figures… PROTON defines about 250 classes and 100 properties Providing coverage of most of the upper-level concepts necessary for semantic annotation, indexing, and retrieval A development of the KIMO ontology A modular architecture, allowing for great flexibility of usage and extension: SYSTEM module - contains a few meta-level primitives (6 classes and 7 properties); introduces the notion of 'entity', which can have aliases; TOP module - the highest, most general, conceptual level, consisting of about 20 classes; UPPER module - over 200 general classes of entities, which often appear in multiple domains.

    6. Ontotext Lab #6/69 16 February, 2012 PROTON Ontology Language The current version of the ontology is encoded in OWL Lite. A few custom entilement rules (axioms) are also defined for usage in tools that support them, for instance: Premise: <xxx, protont:roleHolder, yyy> <xxx, protont:roleIn, zzz> <yyy, rdf:type, protont:Agent> Consequent: <yyy, protont:involvedIn, zzz> Axioms of this sort are built in Sesame (and OWLIM). PROTON is portable to any OWL(Lite)-compliant tools. PROTON can be used without such axioms either.

    7. Ontotext Lab #7/69 16 February, 2012 Other Standards: Relations ADL Feature Type Thesaurus and GNS the backbone of the Location branch; on its turn aligned with the geographic feature designators, of the GNS database of NIMA; PROTON is more coarse-grained, taking about 80 out of 300 types. Dublin Core the basic element set available as properties of protont:InformationResource and protont:Document classes; the resource type vocabulary is mapped to sub-classes of InformationResource. OpenCyc and WordNet– consulted and referred to in glosses. ACE (Automatic Content Extraction) annotation types – covered. FOAF – assure easy mapping (e.g. the Account class was added). DOLCE, EuroWordnet Top, and others – consulted to various extent.

    8. Ontotext Lab #8/69 16 February, 2012 Other Standards: Compliance Other models are not directly imported (for consistency reasons) The mapping of the appropriate primitives is easy, on the basis of a compliant design, and formal notes in the PROTON glosses, which indicate the appropriate mappings. For instance, in PROTON, a protont:inLanguage property is defined as an equivalent of the dc:language element in Dublin Core with a domain protont:InformationResource and a range protont:Language

    9. Ontotext Lab #9/69 16 February, 2012 KIMO: The Predecessor of PROTON (I) Derived from the KIMO ontology: http://www.ontotext.com/kim/kimo.rdfs KIMO used to be “the ontology of KIM”; KIM is a semantic annotation, indexing, and retrieval platform; KIM already migrated to PROTON ontology; mappings of the KIMO URIs to the PROTON URIs available. Major developments of PROTON with respect to KIMO: represented in OWL Lite (KIMO was in RDFS); broken down into modules (System, Top, Upper); KIM-independent – the KIMO-related concepts were relocated to specific, separate modules; some tuning took place in response to the requirements of SEKT (WP5 “Knowledge Access” and the digital library case studies, WP11).

    10. Ontotext Lab #10/69 16 February, 2012 KIMO: The Predecessor of PROTON (II) A part of KIMO proved too KIM Platform-specific, so it was taken out of PROTON and organized into two standalone modules: KIM System Ontology: contains a few meta- or system-level primitives, used by KIM; available at http://www.ontotext.com/kim/2004/12/kimso KIM Lexical Ontology: contains some lexical, resource-related concepts, mostly used to represent lexica required by the KIM information extraction sub-system; available at http://www.ontotext.com/kim/2004/12/kimlo structures the information necessary for information extraction systems (used in some semantic annotation systems); important for such IE systems, whereas not relevant to most of the applications that do not perform any extraction or annotation.

    11. Ontotext Lab #11/69 16 February, 2012 The Architecture of PROTON (I)

    12. Ontotext Lab #12/69 16 February, 2012 The Architecture of PROTON (II) 3 modules (layers) Diacritical modular architecture Allows for easy integration, extension, and remodelling System module subservient, operative, may need to be hard-coded a level that can be also classified as an “application ontology” defines some notions and concepts of a technical nature that are substantial for the operation of any ontology-based software, such as semantic annotation and knowledge access tools Top module – most basic entity types Upper module sub-classes of top classes some branches include much more specialized entities borders with some domain ontology characteristics

    13. Ontotext Lab #13/69 16 February, 2012 protont: Top Module In Brief (I)

    14. Ontotext Lab #14/69 16 February, 2012 Naming Conventions Classes: the label of a class is composed of one or more words, written with capital first letters for each of the words, and without any intervals or alphanumeric symbols between them (in case a class label is a two-word one). Example: protont:InformationResource Properties: the labels of properties follow the same rule, except for the non-capital first letter. Example: protont:locatedIn

    15. Ontotext Lab #15/69 16 February, 2012 protont: Top Module In Brief (II) The most essential ingredient in the PROTON pie The most general classes, required by the SEKT case studies Referred to via the "protont:“ prefix Hierarchy - at the upper-most layer: basic philosophically-reasoned distinctions between entity types: protont:Object – existing entities, as agents, locations, vehicles; protont:Happening – events and situations; protont:Abstract – abstractions that are neither objects nor happenings; respective DOLCE equivalents: Endurant, Perdurant, Abstract. Further specialized by substantially real entity types of general importance: persons, locations, numbers, time, money, roles, employment (job) positions, commercial, government, and other organizations, etc. Characteristic properties (attributes and relations) for featured entity types

    16. Ontotext Lab #16/69 16 February, 2012 protonu: Upper Module In Brief (I)

    17. Ontotext Lab #17/69 16 February, 2012 protonu: Upper Module In Brief (II) Represents the sub-class branches under Top module classes Defines much more specific classes, e.g.: protonu:Mountain, as a specific type of protont:Location; protonu:ResourceCollection as a sub-class of protont:InformationResource Over 200 general classes of entities Upper module classes often appear in multiple domains (e.g. various sorts of organizations, a comprehensive range of locations, etc.) Referred to via the "protonu:“ prefix In the future – can be considered to be “cleaned up” from entities that are too domain-specific; split down further into sub-modules if needed (e.g. the protont:Location branch).

    18. Ontotext Lab #18/69 16 February, 2012 protons: System Module (I)

    19. Ontotext Lab #19/69 16 February, 2012 protons: System Module (II) Referred to via the "protons:“ prefix Classes: protons:Entity represents the root of the ontology the super-class of the Top module (the other branches of System could be considered auxiliary or “system” ones) protons:EntitySource key role – limited support for recording of the provenance of the instance data; its instances are used to separate the trusted (pre-populated) information in the KB, from the one that is extracted automatically; such a distinction is indicated by the protons:generatedBy property of the specific entity; each instance of protons:Entity could be linked to an instance of protons:EntitySource via the protons:generatedBy property;

    20. Ontotext Lab #20/69 16 February, 2012 protons: System Module (III) two sub-classes of protons:EntitySource: protons:Recognized serves to identify a source (like a program or a module) that is able to recognize and generate new entities from a text as part of IE or data mining tasks; those entities - not checked for correctedness, consistency, relevance, etc., therefore not trustable. protons:Trusted may be used to indicate entities, imported from “trusted” sources, like GEOnet Names Server, World Fact Book, GATE/MUSE/KIM gazetteers, but also any other source that may be counted on to provide “trusted” information in terms of its correctedness in some universal sense of “being true”.

    21. Ontotext Lab #21/69 16 February, 2012 protons: System Module (IV) protons:LexicalResource mostly dedicated to encoding of various data (company suffixes like “AG” and “Ltd.”, first names of persons, etc), related to IE and data mining processes. expected to prove useful for use cases within (and outside) the SEKT project. All the branches that used to be located “under” it, except for protons:Alias, were moved to, and are now considered part of, the KIM-specific KIM Lexical Ontology. protons:Alias an important class within this branch; represents the (various) names of the instances of the protons:Entity class.

    22. Ontotext Lab #22/69 16 February, 2012 protons: System Module (V) High-level system- or meta-primitives and properties: protons:description denotes a textual description of an entity, usually a free text in some natural language; as defined for dc:InformationResources; a specialization of rdf:comment in some sense. protons:laconicDescription denotes an extremely short description of an entity; a sub-property of protons:description. protons:generatedBy identifies the party that introduced the entity into the respective knowledge base; relates protons:Entity and protons:EntitySource.

    23. Ontotext Lab #23/69 16 February, 2012 protons: System Module (VI) protons:hasAlias refers an alias of the entity; relates the system classes protons:Entity and protons:Alias protons:hasMainAlias refers the official/most important alias of the entity; a sub-property of protons:hasAlias. protons:systemPrimitive an OWL:AnnotationProperty; used to encode system-specific information for system classes and properties, their instances, and related information, which should not be presented to the end-user.

    24. Ontotext Lab #24/69 16 February, 2012 protons: System Module (VII) protons:transitiveOver an OWL:AnnotationProperty; suggests that one property is transitive with respect to another one, thus making the modelling of a specific, but rather useful modelling pattern, possible; the semantics is defined with the following axiom: (p,transitiveOver,q) (x,p,y) (y,q,z) => (x,p,z)

    25. Ontotext Lab #25/69 16 February, 2012 protont:Object Entities that could be claimed to exist (in some sense of existence). An object can play a certain role in some happenings. Objects could be substantially real (as the Buckingham Palace or a hardcopy book) or substantially imperceptible (say, an electronic document – it only “exists” virtually, one cannot touch it). Properties: protont:hasContactInfo - allows for relations between the protont:Object and protont:ContactInformation protont:isOwnedBy - relates a particular organization to the agents, which are members of that organization

    26. Ontotext Lab #26/69 16 February, 2012 protont:Agent (I) Something, which can show (carry out) an independent action, whether consciously or not. Most animals are considered agents, in most contexts; so are most organizations. According to DOLCE, agents are "objects to which we ascribe intentions, beliefs, and desires". Also denotes any automatic services, including web services and servers.

    27. Ontotext Lab #27/69 16 February, 2012 protont:Agent (II) Properties of protont:Agent: protont:involvedIn protont:isLegalEntity - determines whether a particular protont:Agent is a legal entity or a non-legal one; modelled as a property in order to avoid multiple inheritance of classes and/or multiple classifications of instances protont:partiallyControls - any sort of partial control (incl. ownership) of an agent with respect to an object; an intransitive relation

    28. Ontotext Lab #28/69 16 February, 2012 protont:Person An agent, which is an individual who is a human being (i.e. any living or extinct member of the family Hominidae) Properties: protont:hasPosition - relates a person to a JobPosition protont:hasProfession - relates a person to a Profession protont:hasRelative - relates a person to another one (relative); a many-to-many, bi-directional relationship, with a number of specializations protont:isBossOf - relates a protont:Person to another one, who he is an immediate boss or supervisor of; a many-to-many relationship

    29. Ontotext Lab #29/69 16 February, 2012 protont:Group, protont:Organization protont:Group a group of agents, which is not necessarily organized in any way could be the group of people within a bus or the shareholders of a company protont:Organization set as a sub-class of protont:Group because of this very difference in the presence or the absence of organization of the agents in the group a group, established so that certain known relationships and obligations exist between the members, …. the informal definition is adapted from OpenCyc

    30. Ontotext Lab #30/69 16 February, 2012 protont:Location (I)

    31. Ontotext Lab #31/69 16 February, 2012 protont:Location (II) Usually, a geographic location on Earth, however any sort of 3D regions also fit in this class. The branch contains over 100 sub-classes, aligned with the Alexandria Digital Library Feature Type Thesaurus and GNS Not included as specializations of Location in PROTON: Administrative areas (those are to be classified directly as instances of Location); Territorial waters; Tribal areas.

    32. Ontotext Lab #32/69 16 February, 2012 protont:Location (III) Properties of protont:Location: protont:NIMA_GNS_DSG – the designator of the entity according to the NIMA GeoNames Server. protont:NIMA_GNS_UFI – the Unique Feature Identifier from the NIMA GNS. A number that uniquely identifies the location. protonu:hasUniversity – a relation between protont:Location and protonu:University (a sub-class of protont:Group). protont:latitude – in degrees, minutes, and seconds. protont:longitude – in degrees, minutes, and seconds. protont:subRegionOf - the general part-of relation, which here designates a place between a whole and each of its parts. It has a number of specializations.

    33. Ontotext Lab #33/69 16 February, 2012 protont:Statement A message that is stated or declared. A communication (oral or written) setting forth particulars or facts etc. e.g. "according to his statement he was in London on that day“. Adapted from WordNet. Properties: protont:statedBy: best matches the dc:creator element (an entity primarily responsible for making the content of the resource; examples of a Creator include a person, an organization, or a service); relates to the protont:Agent class; protont:validFrom, protont:validUntil: define the range (period) of validity of the information resource; they align to a great extent to the dc:value element refinement, as defined in Dublin Core.

    34. Ontotext Lab #34/69 16 February, 2012 protont:InformationResource protont:InformationResource denotes an information resource with an identity, as defined in Dublin Core; considered any communication or message that is delivered or produced, taking into account the specific intention of its originator, and also the supposition (and anticipation) for a particular audience or counter-agent in the process of communication (i.e. passive or active feed-back). protont:InformationResource branch designed with tight adherence to the Dublin Core; close in correspondence to what is considered dc:metadata in DC consult documentation for further details

    35. Ontotext Lab #35/69 16 February, 2012 protont:Document protont:Document the information content of any sort of a document; any tangible (material) aspects of a document are ignored; usually a document in free text with no formal structure or semantics; a distinctive type of an information resource: by the aspects of intention and “particularism” of a document, at least in the “general” connotations and understanding about “what a document is" and what cannot be described as a document (e.g. a database is not a document, it is rather a dataset).

    36. Ontotext Lab #36/69 16 February, 2012 protont:Product Covers the general concept of a product model (e.g., Ford T). The instances of this class - not specific instances of the product but just objects. Analogous to FormalProductType in Cyc. A proprietary property: protont:producedBy - a relation between the protont:Product and protont:Company/protont:Agent which produced/produces it. Classes protont:Service and protont:Product have been moved from Upper to Top in order to avoid the dependence of SKULO on the Upper module.

    37. Ontotext Lab #37/69 16 February, 2012 protont:Service Denotes any sort of a service, ranging from scheduled flight or train services to weather forecast information/web service. Many services could be considered instances of protont:Agent. protont:operatedBy a proprietary property a relation between a protont:Service and a protont:Agent (usually an organization), which provides it.

    38. Ontotext Lab #38/69 16 February, 2012 protont:Happening Something that “happens”, or “is happening”. Can be either dynamic – like in "drawing a circle", or static – like in "being a president" or “sitting on a chair”. A happening has a certain temporal positioning (extent) – in the simplest case one, denoted by start and end points in time. Properties: protont:startTime - the starting moment of a happening; protont:endTime - the end point of a happening. Often some entities take part in the happening – i.e. are involved in, or play a certain role in it. To this end, PROTON contains the protont:Role sub-class.

    39. Ontotext Lab #39/69 16 February, 2012 protont:Event Denotes a dynamic event, such as "running", or "a concert". In a binary logical opposition to protont:Situation in terms of the dynamic vs. static nature of the protont:Happening. Inherits the properties of its super-class protont:Happening. Its specializations include sub-classes like protonu:Accident protonu:Meeting protonu:Project protonu:SportEvent protonu:MilitaryConflict protonu:ArtPerformance

    40. Ontotext Lab #40/69 16 February, 2012 protont:Situation Denotes a static event or situation, like "sitting on a chair" or "holding position". Typically, those are temporarily homogeneous, i.e. their nature is not expected/required to change within their duration. Usually, to happen/take place/be true for some periods of time and may or may not have a well-defined space extension. In a binary logical opposition to protont:Event in terms of the dynamic vs. static nature of the protont:Happening. Inherits the properties of its super-class protont:Happening. Its specializations include sub-classes like protont:JobPosition protont:Role

    41. Ontotext Lab #41/69 16 February, 2012 protont:TimeInterval A general time expression, which refers to a particular period of time, an interval. Repeating periods (like the Spring or Christmas) are not time intervals, while specific instances of theirs (like the Spring of 1944) are. considered a specialization of protont:Happening, because this is the place in PROTON for all entities that are largely defined through their temporal extent. Abstract seasons, months, days of the week are represented as sub-classes of the protonu:TemporalAbstraction in the protont:Abstract branch of PROTON. Thus, the general notions of Tuesday and July appear as instances of classes in this branch.

    42. Ontotext Lab #42/69 16 February, 2012 protont:JobPosition (I) protont:JobPosition-s of people within organizations are modelled as a special sort of a situation, i.e. a “static” happening. “It happened that he was a CTO at XYZ Corp between 2001 and 2003.” Properties: protont:holder – relates the position to the protont:Person who holds it. There is an inverse relation to this one, named hasPosition, defined for persons. withinOgranization – the protont:Organization where the position is situated. protont:heldFrom and protont:heldTo – two literal-ranged attributes, to be used to encode the start and end of the period, for which the particular person used to take this position, if known. Those are specializations respectively of the protont:startTime and protont:endTime attributes of protont:Happening.

    43. Ontotext Lab #43/69 16 February, 2012 protont:JobPosition (II) a general temptation (or expectation) to model protont:JobPosition-s with relationships (instead of classes/instances), e.g. <pers1, pos1, org1>. an alternative to the current approach would be some sort of an “n-ary” relation or reification, none of which could be decided on as a better option

    44. Ontotext Lab #44/69 16 February, 2012 protont:JobPosition (III) When using sub-classes of protont:JobPosition, it should be taken into account that such a sub-class should still be instantiated to model a particular position.

    45. Ontotext Lab #45/69 16 February, 2012 protont:Role designates the role of an entity (usually an agent) within/for/during/effecting (intentionally or not) a particular happening. For instance, the role played by a coordinator of a project, or a defendant in a trial, or even an evidence in a trial. Proprietary relations and attributes are: protont:roleHolder – relates the role to the protons:Entity that holds (“plays”) it; protont:roleIn – relates the role to the protont:Happening that conditions and takes on the effect of the role; by an axiom in PROTON, when the protont:roleHolder of a role instance is an instance of a protont:Agent, then this protont:Agent is assumed to be protont:involvedIn the respective protont:Happening

    46. Ontotext Lab #46/69 16 February, 2012 protont:Abstract Denotes an abstraction: i.e. something, which neither happens nor exists, e.g. a number or a chemical compound. Those are usually some symbols invented to refer to general notions. In some cases, when a class has specific, well-definable instances, which however may not be modelled as sub-classes, those instances are included in the knowledge base (e.g. some special instances of the protonu:Profession class - artist, politician, religious person, scientist, sportsman – have been instantiated in the system part of the ontology).

    47. Ontotext Lab #47/69 16 February, 2012 protont:Number any given number, within the meaning that a number is: a concept of quantity derived from zero and units (“every number has a unique position in the sequence"); or a number is a numeral or string of numerals used for identification ("she refused to give them her Social Security number"); or a phone number, etc. protonu:Percent sub-class - a further specialization; denotes a specific percent value and thus it is a quite more specific kind of a number in terms of representation and meaning

    48. Ontotext Lab #48/69 16 February, 2012 protont:ContactInformation Any instance of a particular notation, used to make the contact with an individual or an organization possible.

    49. Ontotext Lab #49/69 16 February, 2012 protont:Language denotes a spoken or written natural language i.e. a human written or spoken language used by a community; opposed to e.g. a computer language (adapted from WordNet). An important property protont:inLanguage concerns protont:Language but not proprietary to this class; relates the protont:InformationResource and protont:Language top module classes; the domain of this property is the protont:InformationResource class; links an information resource to the language of the intellectual content of the resource.

    50. Ontotext Lab #50/69 16 February, 2012 protont:Topic Any sort of a topic or a theme, explicitly defined for classification purposes. As long as any other class or entity can play the role of a topic, the instances of this class are only those concepts, which are defined to serve as topics. The topic class is the natural top-class for linkage of logically informal taxonomies. PROTON does not provide any protont:Topic branches as part of its Upper module layer.

    51. Ontotext Lab #51/69 16 February, 2012 protont:GeneralTerm The “place” for the representation of a general concept with a well-defined (idiomatic) meaning, which can have a set of distinct lexical items (surface realizations) associated with it. Such examples are: F2F; I18N; P2P; B2B; VIP; ASAP; Semantic Web…

    52. Ontotext Lab #52/69 16 February, 2012 Usage and Extension Guidelines (I)

    53. Ontotext Lab #53/69 16 February, 2012 Usage and Extension Guidelines (II) PROTON - a lightweight ontology - not packed with an excessive number of logical axioms (unlike comprehensive upper-level ontologies). built with just a basic subsumption hierarchy and a few axioms The diagram demonstrates the dependencies between the modules of PROTON and other components. depicts possible extension/customization paths; to be interpreted from the bottom level upwards. The PROTON System module the most basic module; to be imported into any ontology or KM tool, developed within the SEKT project.

    54. Ontotext Lab #54/69 16 February, 2012 Usage and Extension Guidelines (III) hardcoded in PROTON, stays as a must in the process of integration of PROTON within the process of development of other, more specialized ontologies, applications, and/or KM tools; the Top and Upper modules - can be well overlooked and substituted, without any “damage”, with alternative ones (or simply modified to a certain extent); the System module is quite basic; provides a starting place for the mapping of the Top and Upper modules of PROTON.

    55. Ontotext Lab #55/69 16 February, 2012 Usage and Extension Guidelines (IV) Top and Upper modules – their employment in the development of KM tools and the construction of other ontologies – not a must, though highly recommended. Recommendations: at least the Top module should be used in any development in the scope of SEKT; however not a necessity; an important consideration - SKULO ontology (SEKT Knowledge Management Upper Level Ontology) is dependent on both the Top and Upper modules of PROTON e.g. skulo:Device is dependent on protont:Product and protont:GeneralTerm; this means that any software/ontology, which depends on SKULO, should also import the Top and Upper modules of PROTON.

    56. Ontotext Lab #56/69 16 February, 2012 Usage and Extension Guidelines (V) Custom Top Ontology and Custom Upper Ontology stand for any other ontologies that may extend or substitute (partially or completely) the proprietary PROTON Top and Upper modules, respectively. Such a replacement not difficult to accomplish; however the PROTON Top and Upper modules should be efficient, consistent, and generic enough to satisfy the needs of the majority of use cases within and beyond the SEKT project. SKULO - an example of an upper-level ontology that extends PROTON with those sets of classes, properties, and relations, which are available to all SEKT applications, irrespective of the application domain, but which are not included in PROTON.

    57. Ontotext Lab #57/69 16 February, 2012 Usage and Extension Guidelines (VI) Application Ontology and Domain Ontology – stand for specific application and/or domain ontologies that will be mapped to PROTON as its extensions, according to the requirements of each particular use case. KIM-specific KIM System Ontology and KIM Lexical Ontology mentioned because of their smooth integration with the predecessor of PROTON in the past – the KIMO ontology; this fact allows for their easy use (and reuse) when (and if) that may prove relevant; the KIM Platform dependent on the System module; one can use KIM via importing the System module and extending it with another domain and/or upper ontology; some tuning is required because the default information extraction module of KIM is coupled with grammars, dependent on the Top and Upper modules.

    58. Ontotext Lab #58/69 16 February, 2012 Technical Info Web site, “physical” address PROTON home Web site: http://proton.semanticweb.org provides up-to-date versions of the PROTON modules; PROTON-related documentation (D1.8.1) KIMO-PROTON mappings Namespaces PROTON System: http://proton.semanticweb.org/2004/12/protons PROTON Top: http://proton.semanticweb.org/2004/12/protont PROTON Upper: http://proton.semanticweb.org/2004/12/protonu the namespaces to stay as defined above as longer as possible, notwithstanding the year and month of production cited in the URL-s. major updates and subsequent changes of the namespaces to be announced and implemented on the PROTON home Web site

    59. Ontotext Lab #59/69 16 February, 2012 PROTON Works Best With … PROTON was extensively tested for correctedness and performance with the OWLIM SAIL of Sesame OWLIM is an OWL Lite In-Memory storage and inference layer (SAIL) for Sesame v. 1.1. OWLIM supports partial reasoning over OWL Lite. It is available from http://www.ontotext.com/downloads/index.html#sesame-sail. This configuration (OWLIM + Sesame 1.1) is used by KIM v. 1.2. This means that high performance and scalability up to millions of statements is guaranteed.

    60. Ontotext Lab #60/69 16 February, 2012 Future Development (I) Refinement of protont:Location branch. such an improvement - triggered by known cases when the current status of some properties may lead to factually untrue inferences, for instance: protont:subRegionOf is an owl:TransitiveProperty; "Diego Garcia" is a protonu:MilitaryAreas located in the Indian Ocean; "Diego Garcia" protont:subRegionOf "USA"; "USA" protont:subRegionOf "North America"; thus, because of protont:subRegionOf transitivity, it may be wrongly assumed that "Diego Garcia“ protont:subRegionOf "North America". Re-modelling of protonu:Employee class into a property. current ‘class’ status may lead to problems with metadata tools; the “employee” status of a person should be considered more a relation between a Person and an Organization (or, in some cases, between two Persons) rather than an instance of an Employee class.

    61. Ontotext Lab #61/69 16 February, 2012 Future Development (II) Slimming down multiple-inheritance of classes multiple-inheritance of classes at the lower levels should be reduced to an adequate minimum; aim: to avoid problematic metadata generation due to potential ambiguity or inconsistent results when processed by metadata tools. Remodelling of protons:EntitySource and protons:generatedBy: current modelling approach - helpful for the SEKT-related applications, projects, and tools; however, it narrows the use of such additional information to instances only, for instance: SEKT WP1/WP3 – might be interested in additional information on concepts/properties; vs. SEKT WP2 – might have essential interest in information on property extensions. protons:Recognized and protons:Trusted subclasses - to be remodelled as owl:AnnotationProperty, since their semantics may appear not clear enough and even arbitrary to a part of the community.

    62. Ontotext Lab #62/69 16 February, 2012 Future Development (III) Remodelling of other System module entities - analogously for protons:LexicalResource, protons:Alias, protons:hasAlias, protons:hasMainAlias, protons:description, and protons:laconicDescription. generally useful conceptualizations for SEKT, but a serious need for a possibility to encode rich lexical information and descriptions not only for instances, but also for concepts, properties, and even property extensions. a possible solution - remodelling as owl:AnnotationProperty.

    63. Ontotext Lab #63/69 16 February, 2012 Future Development (IV) Remodelling of protonu:PublicCompany defined as “a company, which is publicly traded on some stock exchange”; goes too far into detail for an upper-level ontology like PROTON; metadata generation tools based on IE typically experience problems with multiple classifications of instances; quite domain-dependent - to be modelled as a proprietary property of protonu:Company instead of a sub-class. Conversion of protonu:hasUniversity - a more generic property to relate Organizations and Locations.

    64. Ontotext Lab #64/69 16 February, 2012 PROTON Community Process (I) To involve all SEKT partners and workpackages To be set up in the immediate future Purpose: all comments, suggestions, arguments, and issues, relating to PROTON, to be set forth and out and put to discussion in a fairly structured and easily traceable way. Final aim: to resolve such issues for the sake of some consensual ontology improvements. Results and updates – to be reflected in the guidance at certain regular intervals. Technical organization - could involve the use of tools like Wiki and Plone. Another option: to investigate the possible application of the DILIGENT methodology available from University of Karlsruhe, Institute AIFB; http://www.sekt-project.com/internal/sektwikis/diligent/SektTopLevelOntology purpose: the engineering of upper-level ontologies like PROTON.

    65. Ontotext Lab #65/69 16 February, 2012 PROTON Community Process (II) A potential further refinement of the modular architecture of PROTON a currently known issue for discussion within the future community process; reasons: the modular dividing line between the Top and Upper modules allows for a great extent of flexibility; however, in the process of PROTON usage and extension, this border line may prove slightly imprecise or indistinct. possible considerations that certain Top module classes be moved to the Upper module, and/or vice versa e.g. the protont:hasPosition and protont:isBossOf properties; the Upper module may be broken down into smaller sub-modules - e.g. the protont:Location branch.

    66. Ontotext Lab #66/69 16 February, 2012 Conclusions (I) PROTON is a modular, lightweight, upper-level ontology - major assets: domain-independent; compliant with popular metadata standards; provides light-weight logical definitions; ensures an broad coverage of concrete and/or named entities; requires minimal support for general concepts, which ensures the easy extension in this direction; encoded in OWL Lite; contains a minimal set of custom entilement rules (axioms); Current version of PROTON - its very first one. After an extensive discussion within a community process - a final version of PROTON to be released at the end of year 2005 (SEKT deliverable D1.8.2, due M24).

    67. Ontotext Lab #67/69 16 February, 2012 Conclusions (II) Documented in the guidance (D1.8.1 SEKT deliverable), presenting: design rationales and decisions, comments on the alignment to other ontologies and metadata schemata; an introductory exposition of the major layers and branches in PROTON; extensive descriptions of the three modules of PROTON; a discussion on the further development of PROTON, the relations and dependencies between the two ontologies, as well as the design principles observed in the process of designing the modular architecture of PROTON; a general introduction to ontologies, knowledge representation, and ontology languages; a discussion on the scope, coverage, and compliance of PROTON and its relations to, and alignment against, other standards.

    68. Ontotext Lab #68/69 16 February, 2012 Conclusions (III) Due to its unique modular architecture and well-organized subsumption hierarchy, PROTON is a flexible, lightweight upper level ontology; easy to adopt and extend for the purposes of the tools and applications developed within SEKT project; quite all-purpose in the sense that it describes very general concepts like space, time, events, objects, abstractions, etc., which for the most part are independent of a particular problem or domain built with just a basic subsumption hierarchy and a few axioms a general purpose ontology; quite easy to understand and interpret; The maintenance, modification, and/or extention of PROTON would require minor efforts.

    69. 16 February, 2012 PROTON http://proton.semanticweb.org A Base Upper-Level Ontology for the Semantic Web Ivan Terziev, Atanas Kiryakov SEKT Q4 Meeting, Innsbruck, January 17-19, 2005

More Related