Knowledge as CI: Toward Geographic Knowledge Systems Rob Raskin NASA/JPL.

Knowledge as CI: Toward Geographic Knowledge Systems

Rob RaskinNASA/JPL

Data to Information to Knowledge

Data Information Knowledge

Basic Elements Bytes Numbers Models FactsStorage File Database GIS FTC Ontology MindVolume High LowDensity Low HighServices Save Discover Visualize Overlay Infer Understand Predict

Syntax Semantics

What is Spatial Knowledge?

• Spatial facts, relations, meanings• Content with context

Information is summarized or organized data

• Common sense• Shared understanding• Suitable for spatial reasoning/inference• Dynamic, expandable knowledge base• Requires semantics

Ontology: Knowledge as Cyberinfrastructure

• Context Shared understanding of concepts

• Content A dictionary readable both by computers and

humans

• Form A namespace (URL/URI/pointer) containing an

authoritative declaration of a concept

• Authority Anyone… but the best will win out We will vote with our fingers

Enterprise-wide Ontology

• Smart discovery Google does not understand meaning of

terms Neither does ArcGIS

• Smart integration• Knowledge commons

Assist education

Why point to an ontology?

• Enable machine-to-machine communication

• Associate data with its context• Remove ambiguity

Temperature anomaly: relative to what climatological average?

Remotely sensed feature measured how

Spring (season) or Spring (water source)

Use Case: Global Warming Query

Find data which demonstrates global warming at high latitudes during summertime and plot warming rate.

Extract information from the use-case - encode knowledge Translate this into a complete query for data - inference and

integration of data from instruments, indices and models

“Global warming”= Trend of increasing temperature over large spatial scales

“High latitude”= Latitudes > 60 degrees“Summertime”= June-Aug (NH) and Jan-Mar (SH)“Find data”= Locate datasets using catalogs, then access

and read it“Plot warming rate”= Display temperature vs time

Ontology Representations…

• XML• Triples• DBMS• Visual

Ontology Representation: Triples

Subject-Verb-Object representation

• Flood is a WeatherPhenomena• GeoTIFF is a FileFormat• Soil Type is a PhysicalProperty• Pacific Ocean is a Ocean

• Ocean has substance Water• Sensor measures Temperature

Ontology Representation:

Visual

Atmosphere

AtmosphereLayer

Troposphere

Tropopause

Stratosphere

isUpperBoundaryOf isLowerBoundaryOf

subClassOfsubClassOf

partOf

PlanetaryLayer

partOf

3DLayer

subClassOf

upperBoundary=50 km

lowerBoundary=15 km

primarySubstance=“air”

sameAs=“LowerAtmosphere”

Plate tectonics - before

Plate Tectonics Ontology

Ontology of an Organization

• <owl:Class rdf:ID="StormSurge">• <rdfs:subClassOf rdf:resource="#Flood"/>• <rdfs:subClassOf>• <owl:Restriction>• <owl:onProperty rdf:resource="#hasAssociatedEarthRealm"/>• <owl:allValuesFrom rdf:resource="#earthrealm.owl#CoastalRegion"/>• </owl:Restriction>• </rdfs:subClassOf>• </owl:Class>

• <owl:Class rdf:ID="Flood">• <rdfs:subClassOf rdf:resource="#SevereWeatherPhenomena"/>• <rdfs:subClassOf>• <owl:Restriction>• <owl:onProperty rdf:resource="#hasAssociatedEarthRealm"/>• <owl:allValuesFrom rdf:resource="earthrealm.owl#LandSurface"/>• </owl:Restriction>• </rdfs:subClassOf>• <rdfs:subClassOf>• <owl:Restriction>• <owl:onProperty rdf:resource="#hasAssociatedSubstance"/>• <owl:allValuesFrom rdf:resource="substance.owl#LiquidWater"/>• </owl:Restriction>• </rdfs:subClassOf>• </owl:Class>

Ontology Representation: XML

Ontology Languages in XML: RDF and OWL

• Use of standard languages make it easy to extend (specialize) concepts developed by others Consistent with how we learn - incrementally

Knowledge passed down to future generations Continues web paradigm - anyone can be an author

• W3C languages specialize XML Resource Description Formulation (RDF)

Standardizes: class, subclass, property, subproperty Ontology Web Language (OWL)

Standardizes: transitive functions, inverse relations, cardinality, etc.

• Open world assumption Facts not explicitly stated are not assumed to be false In contrast, DBMS world uses closed world assumptions

Tuples are the truth, the whole truth, and nothing but the truth

SWEET

• Semantic Web for Earth and Environmental Terminology (SWEET)

• Upper-level concept space for Earth system science

• Includes concepts of: Numerics Space Time Earth system science Data and information Data and information services

Why use an Upper-Level Ontology?

• Many common concepts used across spatial disciplines (such as properties of the Earth surface) Provides common definitions for terms used in multiple

disciplines or communities Provides common language in support of community and

multidisciplinary activities Provides common “properties” (relations) for tool

developers

• Reduced burden (and barrier to entry) on creators of specialized domain ontologies Only need to create ontologies for incremental knowledge

Non-LivingSubstances

LivingSubstances

PhysicalProcesses

Earth Realm

PhysicalProperties

Time

NaturalPhenomena

Human Activities

Integrative Ontologies

Space

Data

Faceted Ontologies

Units

Numerics

SWEET 1.0 Ontologies

SWEET Spatial Ontologies

• Objects 0-D, 1-D, 2-D, and 3-D

• Coordinate systems• Relations

Above, inside, adjacent, etc.

• Fuzzy concepts “near”

• Spatial statistics• Features

SWEET 2.0 Modular Design

Math, Time, Space

Basic Science

Geoscience Processes

Geophysical Phenomena

Applicationsimportation

• Supports easy extension by domain specialists

• Organized by subject (theoretical to applied)

• Reorganization of classes, but no significant changes to content

• No longer one-to-one relation of facet to file

SWEET 2.0 New Features

• Organized by subject Makes it easy for domain specialists to

add new modules• Smaller, modular ontologies

Unidirectional import• 12 large ontologies -> 100 small

ontologies

SWEET 2.0 Ontologies

AbstractTo Applied

Characteristic Level of Abstractions

• Any concept has a characteristic level of abstraction How theoretical is it?

• If more than one characteristic level… repeat it Words are commonly defined in

dictionaries with multiple meanings

Community Agreements:How to Optimally Use OWL

• Reduce quadruples/quintuples to triples Interval quantities

hasLowerBound, hasUpperBound, hasUnit

• Fuzzy concepts “nearlySameAs” relatedConcept connects siblings or other concepts that

might be closely related similarity matrix provides more precise support for

search results ranking

• Ensembles pdf representations

Collaborative Ontology Development

Ontology Development Tools: CMAP

• Free, downloadable tool for knowledge representation and ontology development

• Visual language with input/export to OWL Supports subset of OWL language

• http://cmap.ihmc.us/coe

Resources

• SWEET http://sweet.jpl.nasa.gov

• Ontology development/sharing site http://PlanetOnt.org