Ontology support
Ontology supportFor the Semantic WebThe big pictureDiagram, page
9html5 xml can be used as a syntactic model for RDF and
DAML/OILRDF, RDF Schema (with data modeling) RDF takes object
specifications and flattens them into triplesDAML/OIL used to
specify the details of UPML componentsUPML architectural
description language for components, adapters, connection
configurationsDAML & OILDAML examples, pages 69 to 77OIL
examples, pages 99OIL constraints 101 to 103Intriguing diagram,
page 113
UPMLDiagram of UPMLs role, page 144Key function: component
markupUPML diagram, page 147 a PSM is a problem solving methodProtg
is a free editor for ontology-related languages, page 160 &
162
Another Big view of the semantic webDiagram, page 173Intriguing
comparison diagram, page 175Extra capabilities of ontologies over
lower level specificationsConsistencyFilling in semantic
detailsInteroperability supportValidation and
verificationConfiguration supportSupport for structured
searchesGeneralization/specialization meta informationInteresting
twist on how databases should be builtOld way page 266New way page
268The smarter DB architecture, page 273What are we adding?Used to
be data, schema, then sql, then transaction manager, then apps,
then UINow we are introducing more metadata? More schema?Or is this
a completely different kind of database?Where data consists of
assertions?
A semantic portalPage 320Both humans and agents can access
semantic portalsBut how do humans interact with a semantic portal
via a browser?Comparison between ontologies and knowledge page
322The idea of extensibility as a critical aspect of the semantic
webNot just new data, not just new metadata, but new inferences as
wellBig picture diagram, page 333Semantic Gadgets conceptMaking
smarts ubiquitousThe Internet of Things and Ambient IntelligenceFor
learning, mobile activities, using remote servicesMobile computing
and mobile-based queriesDevices that can interact with our
devicesMuseum locations and user with sound deviceHand held devices
and grocery store shopping and congnitively disabledSemantic
annotation conceptDiagram page 406Detailed diagram page 415Example
pages 417 and 418We see the use of parallel databases that hold
metadata that is searchableAnd metadata can be applied in a
personalized way to provide specific results to specific usersSee
page 420..Task-achieving agents notionDiagram, page 434Kinds of
tasksAutomated planningComputer-supported cooperative
workMulti-agent mixed-initiative planningWorkflow supportExample
diagram, page 442This is a common way of viewing the new webSmart
agents replace browsersA concrete component: SPARQLQuery language
modeled after SQLIt can walk through semantic websites and across
semantic websitesSPARQL thus creates new knowledge by creating
inferences that can cross website boundaries
From -
http://www.cambridgesemantics.com/2008/09/sparql-by-example/A
SPARQL query comprises, in order:Prefix declarations, for
abbreviating URIsDataset definition, stating what RDF graph(s) are
being queriedA result clause, identifying what information to
return from the queryThe query pattern, specifying what to query
for in the underlying datasetQuery modifiers, slicing, ordering,
and otherwise rearranging query resultsWhat can sparql do?It can
extend an ontology by adding new inferences as assertionsRetrieve
triples that describe somethingAsk true or false questions based on
assertions
Another view: The open semantic frameworkLayered
architectureModular softwareIt is part of a four component
approach:SoftwareStructureDocumentationMethods
GoalsLeverage existing data and appsBuild and validate
incrementallyUse open software, standards, protocolsLink dataUse
RDF as a unifying data modelAddress high level IT management
issuesAssumptions and techniquesUse URIs to identify informationAll
data is equal text, media, relational dbs
Big picture
from:http://openstructs.org/open-semantic-framework/overview
layersExisting assetsDatabases of all kindsWeb
pagesDocumentsInformation Transformation (scones/irON)Extraction of
data and metadataScones subject concept or named
entitiesConversions via irON (instance record Object Notation)
Layers continuedstructWSF layerThe workhorseWeb services
frameworkProvides a common interface layer by which existing info
assets can be mediatedInclude CRUD, browse, search, export, import
primitivesSupports sparql Rights and permissions controlsEach
structWSF instance has a unique Web address that allow easy
use/reuse and reconfigurationLayers continuedSemantic Components
layerTakes computed results generated via queries from one or more
structWSF instances and presents data visually using semantic
componentsComponents includeFilterTabular templatesBar, pie, other
chartsRelationship browserAnnotator
Layers continuedOntologies layerContent Management System layer
(conStruct)ThinEndpointsPortalsCollaborative environmentsMedia
richThe big picture: web-oriented
Major goal: domain specific instances
Is it redundant?
Hmm you can download
ithttp://techwiki.openstructs.org/index.php/Open_Semantic_Framework_InstallerAnother
view of
ontologies:http://www.cems.uwe.ac.uk/amrc/seeds/ModellingSemanticWeb.htm
dbpedia:http://www.hewettresearch.com/svcc2009/
What is it?DBpedia isa community effort toextract structured
information from Wikipedia andtomake this information available
onthe Web. DBpedia allows youtoask sophisticated queries against
Wikipedia, andtolink other data sets onthe WebtoWikipedia data.
Wehope this will make iteasier forthe amazing amount ofinformation
inWikipedia tobe used innew andinteresting ways, andthat itmight
inspire newmechanisms fornavigating, linking andimproving
theencyclopaedia itself.
Facts:http://dbpedia.org/AboutTheDBpedia knowledge base
currently describes more than 3.64 million things416,000 persons,
526,000 places, 106,000 music albums, 60,000 films, 17,500 video
games, 169,000 organisations, 183,000 species and5,400
diseasesTheDBpedia knowledge base allows youtoask quite surprising
queries against Wikipedia, forinstance Give meall cities inNew
Jersey with more than 10,000 inhabitants orGive meall Italian
musicians from the18th centuryThe DBpedia data setisinterlinked
with various other datasets onthe Web.A possible application of
semantic web technology: citrus & moreWork with Brad ParksThe
HLB disease caused by a bacteriaSpread by an insect called the
Asian citrus psyllidAttacks all citrus treesHas infected 40% of
trees in Florida, the largest orange producing state in the USHas
been found in Florida and Arizona, insect but not the bacteria in
CaliforniaHas heavily wiped out citrus orchards in Brazil (largest
orange producer in the world) and MexicoIts too late for Florida
& since there is no treatment, tracking does littleBut lots of
pathogens and disease vectors can be tracked and modeledDetectors
in the field (DNA fingerprinting, organic chemical sensors, heat,
imaging)Volunteers on the ground who are connected
MorePossible applicationsFood born disease trackingInfectious
disease trackingOther technologyCoordination of testers live in
fieldApplication of models to mathematically similar situations
Citrus and more, continuedInformation collection and
aggregationIntegration of heterogeneous forms of
informationInternet of things: sensors and people (sorry)Ambient
intelligence (sensors have onboard computers and cellular
connectivity devices)Automatic collection of data into multiple
sites and searched automatically via softwareAutomatic delivery of
information aggregation and analysis resultsAutomatic creation of
dynamic models
