CROC CROC — a — a Representational Representational Ontology for Ontology for ConceptsConcepts
ContentsContents
IntroductionIntroduction Semantic WebSemantic Web ConceptuologyConceptuology LanguageLanguage
CROC CROC — a Representational Ontology — a Representational Ontology for Conceptsfor Concepts
Semantic WebSemantic Web
Making Web content understandable for Making Web content understandable for intelligent agentsintelligent agents
RDF/RDFS/OWL ontologies (state of art) RDF/RDFS/OWL ontologies (state of art) that that define classesdefine classes
The The interoperability probleminteroperability problem: how to : how to merge different world-views?merge different world-views?
ClassificationClassification
Different Different classifications: classifications: “world-views”“world-views”
Classification needs Classification needs identificationidentification
Communication (I)Communication (I)
Communication:Communication: (1) expresses using (1) expresses using
symbolssymbols (2) reads what is (2) reads what is
expressedexpressed
Interoperability Interoperability problemproblem when one when one doesn’t know the doesn’t know the symbolssymbols
CYC: a shared CYC: a shared classification?classification?
CYC.com: developing one big CYC.com: developing one big classificationclassification
One world-viewOne world-view ““not soon” or never completenot soon” or never complete agents have own interests and pick up other agents have own interests and pick up other
ideas (“autonomy”)ideas (“autonomy”) conceptionsconceptions may be different from agent to may be different from agent to
agentagent
Mapping world-views?Mapping world-views?
Should we map Should we map classifications to classifications to solve the solve the interoperability interoperability problem?problem?
Rather: think about Rather: think about the identification the identification mechanism (for a mechanism (for a Semantic Web!Semantic Web!).).
Communication (II)Communication (II)
Communication:Communication: (1) (1) representsrepresents (2) identifies and (2) identifies and
classifiesclassifies
Problem when the Problem when the receiving agent receiving agent cannot identify the cannot identify the representationrepresentation
Identification: Identification: conceptuologyconceptuology
A concept =A concept = (fuzzy / partial) definition?(fuzzy / partial) definition? prototyping?prototyping? an ability to reidentify for a purpose an ability to reidentify for a purpose
[1:Millikan, On Clear and Confused Ideas: [1:Millikan, On Clear and Confused Ideas: An Essay on Substance Concepts]An Essay on Substance Concepts]
Most concepts are not classesMost concepts are not classes
Concept for dogsConcept for dogs
http://en.wikipedia.org/wiki/Image:Pupppppy.jpg
1 2
Common senseCommon sense
Computers usually don’t have much Computers usually don’t have much common sense: they are deaf, blind, common sense: they are deaf, blind, tasteless, touchless, etc.tasteless, touchless, etc.
Do they need it for having concepts?Do they need it for having concepts?
LanguageLanguage
Same concepts, different Same concepts, different conceptionsconceptions Having concepts entirely through Having concepts entirely through
languagelanguage
“It is common [to] have a substance concept entirely through the medium of language. It is possible to have it, that is, while lacking any ability to recognize the substance in the flesh.” [1, Ch. 6][1, Ch. 6]
CROC CROC — a Representational — a Representational Ontology for Concepts (I)Ontology for Concepts (I)
Lexical representations for conceptsLexical representations for concepts Concepts have Concepts have namesnames (so can be shared by (so can be shared by
language)language) Where the name fails, CROC uses induction Where the name fails, CROC uses induction
or deduction using the related knowledge to or deduction using the related knowledge to the conceptthe concept Representation, using other conceptsRepresentation, using other concepts DescriptionsDescriptions instead of instead of definitionsdefinitions
Examples (I)Examples (I)
A: “Swans are white.”OWL B: (OK, I’ll take that into the class
definition.)CROC B: (OK, nice to know.)A: “There is a black swan.”CROC B: (OK, nice to know.)OWL B: (Error in [1], or unalignable classes for
“swan”.)
CROC CROC — a Representational — a Representational Ontology for Concepts (II)Ontology for Concepts (II)
Concepts for every unit of representationConcepts for every unit of representation SubjectsSubjects, subdivided in Kinds (like ‘a dog’), , subdivided in Kinds (like ‘a dog’),
Individuals (like ‘Oscar’), and Stuffs (like Individuals (like ‘Oscar’), and Stuffs (like ‘gold’)‘gold’) SubstancesSubstances Properties Properties (like ‘colour’)(like ‘colour’) Happenings Happenings (events, situations)(events, situations)
Predicates Predicates (like ‘poor’, ‘eager’)(like ‘poor’, ‘eager’) Relations Relations (like ‘of’, ‘in’, ‘at’)(like ‘of’, ‘in’, ‘at’)
CROC CROC — a Representational — a Representational Ontology for Concepts (III)Ontology for Concepts (III)
Abilities to gather, store and query Abilities to gather, store and query representational information for representational information for reidentificationreidentification Storage of statements (happenings) about Storage of statements (happenings) about
conceptsconcepts Subject templatesSubject templates to gather information to gather information Semantical tableaux for reasoning about Semantical tableaux for reasoning about
statementsstatements
Examples (II)Examples (II)
A: “I like Cicero’s De Oratore.”B: (I don’t know that word.) “Cicero??”A: (I will answer what I know is relevant for
humans.) “Cicero is a human. He was born in Arpinum.”
B: (I have other relevant questions about humans.) “Where did he live?”
A: “In Rome.”
Examples (III)Examples (III)
(continued)
B: (I see someone matches all inductive properties.) “Cicero is Marcus Tullius?”
A: “Yes.”
B: (I will merge the two concepts.)
CROC CROC — a Representational — a Representational Ontology for Concepts (IV)Ontology for Concepts (IV)
Our goal is not primarily knowledge Our goal is not primarily knowledge representation, but agent communication representation, but agent communication and understandingand understanding
Agents have their own conceptuologyAgents have their own conceptuology No need for division of linguistic labour No need for division of linguistic labour
(where only experts ‘own’ the concept)(where only experts ‘own’ the concept) Private concepts and conceptions are Private concepts and conceptions are
welcome (“autonomy”)welcome (“autonomy”) Easy learning of new conceptsEasy learning of new concepts
ConclusionsConclusions
Identification by name will be able to Identification by name will be able to solve the interoperability problem (for a solve the interoperability problem (for a great deal)great deal) concepts for every part of the representationconcepts for every part of the representation agents can have own conceptuologiesagents can have own conceptuologies
Concepts may be grounded entirely in Concepts may be grounded entirely in lexical representationslexical representations
Future workFuture work
Higher-order reasoning: about what other Higher-order reasoning: about what other agents believe, etc.agents believe, etc.
A temporal logic for reasoning with statementsA temporal logic for reasoning with statements Integrating classification systems (efficient Integrating classification systems (efficient
knowledge representation)knowledge representation) The language-thought partnership [Millikan, The language-thought partnership [Millikan,
Language: A Biological Model, Ch. 5]Language: A Biological Model, Ch. 5]
Thank you for your Thank you for your attentionattention
http://sourceforge.net/projects/croc
Top Related