Uni Bremen > FB10 > bremen ontology research group

John Bateman's ontology portal...

reference ontologies
graphical history (in progress)
ontology-based standards
(in progress)


This page is a collection of starting points for information on ontologies gathered together for ease of reference for our own ontology-related projects. It is made available as is in case it is of use to anyone else. It does not contain information about conferences, meetings, journals, email groups, and the like. This page stays reasonably general, but since our main concern is with natural language processing, there is a bias.

The explosion of ontologies in the last couple of years also makes any expectation of completeness impossible to fulfill. What should be found here are Upper Ontologies and Generic Ontology components that are the most persuasive currently on offer. There are many many many more, either for limited domains or more shallow in their modelling approach. For further collections, see: the DAML ontology library and SchemaWeb.


I1: OntoSpace

These are links to our current ontology-centered project, concerning ontologies for spatial information and natural language processing

FOIS 2006
FOIS 2004

"Ontology" is a very fashionable term. One can find it being used in very diverse contexts. Ontologies are currently being investigated, for example, in order to transform the World-Wide Web into a far richer semantically based construct. Ontologies are also increasingly refered to for solving problems concerned with organising information in practical domains, such as medicine, geographical information systems, car parts, and many more.

Although this page gives a general overview of current activities in the area, I am particularly concerned with the use of ontologies for linguistic purposes: both organising linguistic information and for relating linguistic information to other kinds of information. For this reason, the linguistic perspective is singled out for development here more than others. WordNets are mentioned, but are not central here as I consider them to be mostly about lexical representation and not ontology; their possible 'interface role' is dealt with separately; similarly for FrameNets  although in both cases there is probably going to be a lot more action in combining these with ontologies, some of which is already reported on below. There is also a clear overlap and lots of interesting action in the Semantic Web effort, several links on this page lead in this direction.

Extensions and comments very welcome.


(these lists claim utility not exhaustiveness!!!)

Ontologies sorted by motivation
  Grammatical GUM
    Penman Upper Model
  Lexical EuroWordNet
    Meaning project Top-Level Ontology (building on EuroWordNet)
    FrameNet (grapher)
  Semantic MikroKosmos
    Sumo (modules)
  Formal Publication list of Nicola Guarino's new Applied Ontology institute (Nov 2002)
    Dolce (Descriptive Ontology for
Linguistic and Cognitive Engineering)
    Sowa's top ontology (best viewed with IE)
    BFO (Basic formal ontology: IFOMIS)
    Smith at Buffalo
    ABC ontology: interoperability between ontologies
    OntoClean Top (OCT)
  Mixed Suo
    Semantic Web
    ANSI Ad Hoc Group on Ontology
    the minimalist upper ontology (gist)

The Multi-Source-Ontology (MSO)

Successor? : A Top Level Ontology from Philippe Martin

    SmartWeb Integrated Ontology (SWIntO)
Ontologies sorted by domain of application
Languages + Linguistics
  Information re-use and maintainability
  General Knowledge
  Process plant, engineering and science
  Software systems Unified Modeling Language (UML)
  Ontology (i.e, upper ontologies and organizations of ontologies)

IFF IFF(home) IFF (work in progress page)

Sumo (modules)
GoL (General Ontology Language)
    Sowa's top ontology (best viewed with IE)
Representations and representational issues
Approaches to relating linguistic and non-linguistic ontologies and mapping mechanisms (prospectively) useful for this task
  Subsumption: domain knowledge subordinated to linguistic ontology: Penman Upper Model, GUM Motivation described in, e.g., Bateman 1990.
  Semantic rerendering: Pustejovsky et al. (2002). Relates UMLS medical information to medical texts.
  Dynamic on-the-fly domain construction according to text genre (Bateman&Teich, 1995).
  Flexical graph matching with CGs (Nicolov, 1999).
  Report on using Cyc for NLP by Mahesh et al. (MCCS-96-302)
  OntoMorph: ' a powerful rule language to represent complex syntactic transformations'
  A general approach to semantic translation across ontologies is the BUSTER project: whether and to what extent this can be applied to the linguistic problem is still to be investigated.
Ontology editing and maintainance tools
    Protégé (continuing development)
    The Sigma Knowledge Engineering tool (as used for SUMO)
    OilEd (just what it says...)
    SRI Ontology construction kit (pdf report)
    TRIPLE (rdf, query, classification, ...)
    OpenKnoME (GRAIL-based ontology development environment of GALEN)
    KSL's Chimaera
    HP's Jena system for semantic web development
    Heterogeneous Tool Set (HETS)
  Bishop John Wilkins (1600). Also the top of Wilkins' hierarchy expressed in SUO-KIF.
  Fritz Lehmann has been compiling a long list of ontologies through the ages for many years (last update appears to be 1996)
  Linneaen Taxonomies (Linneaus: 1707-1778)
Conversion technology
  -> Notes...  



What is an ontology?  

The usual definition that is offered in the context of AI and knowledge representation is the following by Tom Gruber:

"An ontology is an explicit specification of a conceptualization. The term is borrowed from philosophy, where an Ontology is a systematic account of Existence. For AI systems, what "exists" is that which can be represented. When the knowledge of a domain is represented in a declarative formalism, the set of objects that can be represented is called the universe of discourse. This set of objects, and the describable relationships among them, are reflected in the representational vocabulary with which a knowledge-based program represents knowledge. Thus, in the context of AI, we can describe the ontology of a program by defining a set of representational terms. In such an ontology, definitions associate the names of entities in the universe of discourse (e.g., classes, relations, functions, or other objects) with human-readable text describing what the names mean, and formal axioms that constrain the interpretation and well-formed use of these terms. Formally, an ontology is the statement of a logical theory."

Cf. T. R. Gruber. A translation approach to portable ontologies. Knowledge Acquisition, 5(2):199-220, 1993. Available on line.

Cf. also John Sowa's (2001) useful introduction to ontologies as such as well as the application-oriented lead-in from OntologyWorks Inc.

And the W3C take on the matter.

A succinct and far reaching position statement concerning the relationship between traditional philosophical ontology and some modern uses of the term is given in Barry Smith's paper 'Ontology: Philosophical and Computational'.

What is an ontology for?  
    An ontology provides an explicit representation for structuring all the entities and relationships that are to be talked about in some domain. Such organisations are expected to be essential for achieving inter-operability between different kinds of knowledge, as they explicitly define the abstract forms and schemes that all knowledge in that domain will take. They also provide ways of ensure more consistent "modelling styles".  
How do I make one?  

There are several ways in which one can motivate the contents of an ontology. The extent to which efforts with differing motivations can converge on a single ontology-organisation and list of contents is still unclear. It is more likely that different motivations will result in different ontologies.

The main kinds of motivation can be listed as: formal-philosophical, logical-semantic and linguistic. These can be mixed to greater or lesser degrees; in particular, almost all motivations offered for ontologies draw on linguistic evidence somewhere  that is, they consider how particular linguistic items behave in their reasoning about what belongs in an ontology and where. If an ontology is not explicitly aiming at being a linguistic ontology, it is unclear to what extent this kind of argument is justified. I have written a detailed description of the different kinds of ontology motivations relevant when considering Natural Language Processing which can be downloaded in several formats here. Some different motivations are listed in the tables on the right.

The Protégé team offer an Ontology Development Guide that now probably replaces these others: A simple 1-2-3 to get started offered by Tom Gruber here, and Noy and McGuiness's (2001) "Ontology development 101".

For a much more detailed consideration of ontology construction in general, and for a variety of purposes, see the papers:

A good indication of some of the arguments and steps that can (ought?) be gone through when evaluating and critiquing an ontology using the OntoClean methodology is given in this talk by Chris Welty. An in-progress discussion by the group of Nicola Guarino is here.

A paper reviewing metholodogies for ontology construction by Fernández López can be read here (from the IJCAI99 workshop on ontologies and problem solving methods).

A good brief indication of practical issues within the current ontology scene is Michael Denny's (2002) Ontology Building: A Survey of Editing Tools

More advanced methods and descriptions...

Recent publications promoting ontologies  
    CACM Special Issue (Feb 2002)  
    A slide-show by Richard Fikes  
    Pease, A., (2001), Evaluation of Intelligent Systems: The High Performance Knowledge Bases and IEEE Standard Upper Ontology Projects, invited position paper, in Proceedings of the 2001 workshop on Measuring Performance and Intelligence Of Intelligent Systems (PERMIS'2001).
    Overheads of a Panel discussion at: KR-2002 : Are upper level ontologies worth the effort? with Nicola Guarino, Brandon Bennett, Jim Hendler, and Alan Rector.(Download or requires Explorer)

Towards the Semantic Web: Ontology-driven Knowledge Management. Edited by John Davies, Dieter Fensel & Frank van Harmelen (to appear, Dec 2002)

The following claims to be the "first semantic web site":Mindswap, website generated from RDF+ontologies

    Pease, A., Niles, I., (2002) IEEE Standard Upper Ontology: A Progress Report, Knowledge Engineering Review, Special Issue on Ontologies and Agents, Vol 17.  
Some "reference" ontology attempts  

Note: some of the ontologies below can also be inspected from the OntoMap portal. To do this, you must register as a user, but this is free. Documentation for the OntoMap portal can be downloaded here.The OntoWeb portal is also very attractive and contains a substantial quantity of ongoing development.

The beginnings of a graphical view can be found here.

    (D)ARPA Knowledge Sharing Initiative  
    ANSI Ad Hoc Group: "The most recent meeting of the Ad Hoc Group on Ontology Standards [1997] adopted the upper level ontology resulting from merging content from Cyc, Pangloss, WordNet, and Penman. This draft reference ontology is now available for inspection in several formats, as well as on the web from Cycorp and the KSL Ontolingua Server." (KIF, Cyc)  
    DARPA High Performance Knowledge Bases (HPKB) initiative. This overview article appeared in the AI magazine. 1998, Vol. 19, No. 4, pp.25-49. The initiative ended 1997.  
    EuroWordNet Online (online access currently appears limited to Spanish, Catalan and Basque) Further development and access will be pursued in the MEANING project  
    GermaNet: a WordNet in the spirit of EuroWordNet for German.  
    The SUMO (Suggested Upper Merged Ontology) of the SUO working group. Online search DAML version. Teknowlege have also made a graph of the entire thing and give an architecture paper here. I have also put an extract showing the module structure here. But the most uptodate information is on Adam Pease's ontology page here.  
    SENSUS, a 70,000-node terminology taxonomy, as a framework into which additional knowledge can be placed. SENSUS is an extension and reorganization of WordNet (built at Princeton University by George Miller, Christiane Fellbaum, and colleagues); at the top level, we have added nodes from the Penman Upper Model, and the major branches of WordNet have been rearranged to fit. (some is browseable here)  
    OpenCyc: the free extract of the Cyc ontology, represented in CycL. This includes the upper parts of Cyc's conception of a "human consensus reality". The top portion of the hierarchy is also available here in DAML form.  
    GUM: Generalized Upper Model, a grammatically motivated linguistic upper ontology used for interfacing with NLP components.  
    BFO (Basic Formal Ontology): latest development from Barry Smith and colleagues.  
    FrameNet: an attempt to build up a frame hierarchy, where the frames are essentially motivated by lexicographic concerns and corpus instances.  
    CoreLex: "126 underspecified semantic types, the 317 systematic polysemous classes they correspond to and all 39,937 noun instances for each class as derived from WordNet 1.5"
Common Comparison/Combination Tasks  
  Ontology development appears to go in cycles whereby at irregularly reoccuring intervals there are attempts to merge or combine what has been achieved so far. This list contains some of the most significant or interesting of these efforts, loosely chronologically...  
  Upper Cyc and EuroWordNet: Atanas Kiryakov, Kiril Iv.Simov (2000)  
  Penman Upper Model represented in Conceptual Graphs (1995)  
  SUMO to WordNet... (WN-verbs->SUMO WN adjectives->SUMO WN adverbs->SUMO, Rev 1.11 and 1.12, 2002).
Mappings for WN2.0 have now also been added and are available at http://cvs.sourceforge.net/viewcvs.py/sigmakee/KBs/WordNetMappings/.
  SUMO to FrameNet... in progress; paper reporting on this work to be presented at FOIS2006.  
  Wordnet (extracts) in RDF form (now also in OWL)  
  SDTS (Spatial Data Transfer Standard) to Wordnet: Rodrìguez (2000) 'Assessing semantic similarity among spatial entity classes' PhD University of Maine.  
  OntoClean (OCT) to WordNetTOP...
  • WordNet aligned to DOLCE: OntoWordNet (Aldo Gangemi: email)

Levin's English Verb Class Alternations (EVCA)
WordNet and EVCA (paper)
WordNet, EVCA extended and represented as Dorr's Lexical-Conceptual Structure (LCS).
VerbNet and unified combinations (Unified Verb Index)

  Phillipe Martin's Multi-Source Ontology is an attempt to combine most available ontologies into a single knowledge base: well worth a look!
(-> MSO)
  The current attempt of the ONTAC working group to combine OpenCyc, SUMO and DOLCE (->wiki)  
Some other extensive lists, not all of them up to date or using links that are up to date unfortunately