Download presentation
Presentation is loading. Please wait.
Published byIsabel King Modified over 9 years ago
1
Ontologie e applicazioni Marco Brandizi gives_presentation Dec 6, 2005 has_date has_context works_in part_of Dottorato in Informatica XIX Ciclo
2
Outline What an ontology is Informal idea Applications Formal definition Working with ontologies: Description Logics Ontologies and Semantic Web The idea of Semantic Web OWL An application to microarray data management Conclusions and future work
3
Ontologies
4
Issues (relevant in Computer Science, from [Wikipedia]) What is existence? What physical objects? What are essential attributes of objects? Being an apple and being red What constitutes the identity of an object? Categories: Mind: a real entity or a set of states? Physical objects and substances Relationships, type of relations “finger part-of body”? “embryo part-of child-development”?
5
Ontologies and Computer Science From [Mc Guinness]
6
Ontologies and Computer Science From [Mc Guinnes], required properties: Finite controlled (extensible) vocabulary Unambiguous interpretation of classes and term relationships Strict hierarchical subclass relationships between classes Typical, although not required Property specification on a per-class basis Individual inclusion in the ontology Value restriction specification on a per-class basis Desirable: Specification of disjoint classes Specification of arbitrary logical relationships between terms Distinguished relationships such as inverse and part-whole
7
Toy ontology example
8
Example of queries “nails is-part-of elephant”? yes, because: “has-legs” and “has-nails” are particular “has-parts” transitive relation “is-part-of” is the inverse of “has-parts”
9
Example of queries “giraffe is-eaten-by lion”? yes, because: giraffe is-a herbivore => is-a animal lion is-a carnivore => lion eats animal => lion eats giraffe “is-eaten” is the inverse of “eats”
10
Uses (of ontologies and “ontologies”) Reasoning (as in query examples) Consistency checking DUMBO is-a Kenian-Elephant AND DUMBO is-a Lion is wrong Interoperability Two applications may talk together Browsing Searching Sense disambiguation, synonyms, subsumption Completion and interaction Patient is a man => is a male and pregnant is not a cause of its symptoms Natural Language Processing
11
Defining ontologies [Gruber 93]: “An ontology is an explicit specification of a conceptualization” (Variant: it is a shared explicit spec....) Problems [Guarino, 98]: The specification depends by the language which is used for the specification A conceptualization is a more abstract entity, w.r.t. a formal specification which is expressed with a given language
12
What is a conceptualization Conceptualization: the formal structure of reality as perceived and organized by an agent, independently of: the vocabulary used (i.e., the language used) the actual occurence of a specific situation Different situations involving the same objects, described by different vocabularies, may share the same conceptualization. apple mela same conceptualization LILI LELE From [Guarino Tut1]
13
Conceptualizations and ontologies on(x,y) expresses the general idea of on, it has subsets in possible worlds it has restrictions that excludes “impossible” cases for all worlds W1: on_1(a,d) is admitted W2: on_2(a,d) is not admitted here
14
Conceptualizations and ontologies a a a b a Excluded from conceptualization => Impossible in all worlds c
15
From [Guarino Tut1]
18
Formal definition of the ontologies An ontology O for a language L (a set of axioms added to L)approximates a conceptualization C: If there exist a onto. commitment K such that: the intended models of L according to K are included in the models of O An Ontology O commits to C if: has been designed with the purpose of characterizing C Approximates C A language L commits to an ontology O if it commits to some conceptualization C such that O agrees on C
19
Formal definition of the ontologies From [Guarino Tut2]
20
Applying ontologies From [Guarino 98]
21
Designing ontologies Formalized in [Guarino, 94] Apple(A): Apple is a substantial sortal unary predicate Defines A in such a way that a is no longer a apple if ~Apple(A) Red(A): Red is not a sortal, is a characterizing property, if ~Red(A) A is still an Apple Formalized in [Gangemi et al. 2001]: Take account of general, upper level, ontological relations P1)Part(x,x)P2) Part(x,y) and Part(y,x)->x=y P3)Part(x,y) and Part(y,z) -> Part(x,z) Similar work in [ISMB 04] Considers the basic relations and properties that occurs in Biological ontologies
22
Designing ontologies Similar work in [ISMB 04], the OBO relational ontology Considers the basic relations and properties that occurs in Biological ontologies Continuants: entities which endure or continue to exist, changing over time. fetus and embryo are continuants, related one each another by transformation relation Both participate to the Process named human_being_development
23
Outline What an ontology is Informal idea Applications Formal definition Working with ontologies: Description Logics Ontologies and Semantic Web The idea of Semantic Web OWL An application to microarray data management Conclusions and future work
24
Formalizing Ontologies: Logics From [Franconi Tut]
27
1) The entitites which have the Role R with range R being in C 2) The entities that have some y linked to them, as target of role R 1) 2)
28
From [Franconi Tut]
29
DL and computability From [Franconi Tut]
30
DL and computability From [Franconi Tut]
31
Outline What an ontology is Informal idea Applications Formal definition Working with ontologies: Description Logics Ontologies and Semantic Web The idea of Semantic Web OWL An application to microarray data management Conclusions and future work
32
Applications: Semantic Web Web was designed for humans Knowledge on the web is unstructured and meaning is not explicitly represented. It is not machine-readable We would need: Metadata Languages for Data Exchange (XML) Linking Web resources and representing the links meaning (RDF) More advanced representation of Semantics (OWL) Retrieving metadata (SPARQL, SWIRL) Reasoning/Inferencing with metadata (OWL + Rules + Logics)
33
Applications: Semantic Web “The Semantic Web is an extension of the current web in which information is given well-defined meaning, better enabling computers and people to work in cooperation.” Tim Berners-Lee, James Hendler, Ora Lassila, The Semantic WebThe Semantic Web, Scientific American, May 2001
34
(http://owl.mindswap.org)
35
RDF Statements is basic simple structure (Like in Sem Networks) May be represented in XML (as well as in N3) Subject/Property(or verb)/Object may reference public resources, identified by URIs (like Web pages) A given resource may flexibly be annotated by many statements Meaning of statements may be defined by other statements and higher level languages (which still are RDF representable) May be queried with simple query languages (SPARQL, RQL)
36
RDF example From [RDF Nature 05]
37
RDF Example From [RDF Nature 05]
38
RDF Schema (RDF-S) Allow to define simple “ontologies” by means of triples Pro Easy to use Efficiently computable Cons Not enough expressive in several cases: rdfs:range doesn't allow scope management No disjointness of classes No boolean combinations of classes Cannot define a property as transitive/unique/inverse
39
OWL SHIQ Description Logics plus other tools (ex.: imports) Several versions of DL Lite – Fastest, least expressive DL – Computable and expressive, slower Full – Undecidable, most expressiveness XML format Built on top of RDF
40
OWL version 1.2 - A "final stage" that contains all constructs required en
41
OWL BaseEspessa BaseEspessa The inverse property tree to hasIngredient... -
42
Outline What an ontology is Informal idea Applications Formal definition Working with ontologies: Description Logics Ontologies and Semantic Web The idea of Semantic Web OWL An application to microarray data management Conclusions and future work
43
DNADNA genegene mRNA protein Genes Machine Cell/Life
44
Microarray Data
45
MIAME Experiment Modelling From [RDF Nature 05]
46
The mA Experiments Cycle
47
“Closing the loop”
48
What we need to model
51
Microarrays Annotation Ontology Microarray entities Annotation entities
52
Microarrays Annotation Ontology Annotation (source, target, child, parent, rank)
53
Microarrays Annotation Ontology
55
Outline What an ontology is Informal idea Applications Formal definition Working with ontologies: Description Logics Ontologies and Semantic Web The idea of Semantic Web OWL An application to microarray data management Conclusions and future work
56
mA-Anno Ontology: How could it evolve? MGED-2/FuGE Development of some simple tool Methodologies and ontological distinctions Annotation could be an independent ontology Theory of annotations? Named Graphs [NG, NG1] C-OWL [Bouquet et. al.] Context and situations [Gangemi et. al. 2004]
57
mA-Anno Ontology: How could it evolve? From [NG1] b0 skill(Peter, Programming R. Stallmann FSF e-macs getxtext assertedBy authority affiliation
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.