
b08eeab04d6c30c56f53d69fd741580b.ppt
- Количество слайдов: 42
Ontology Collaboration Tools and Services Federation Of Earth Science Information Partners (ESIP) Summer Meeting, July 22, 2010 Knoxville, TN Bruce Bargmeyer Lawrence Berkeley National Laboratory Tel: +1 510 -495 -2905 bebargmeyer@lbl. gov
Topics F Eye on Earth Summit, Abu Dhabi u Technology Infrastructure Workgroup Technology vision for EOE Summit/UNEP F Ontology and voi. D tools F Some practical examples F
Technical Infrastructure An Eye on Earth Summit will be held in Abu Dhabi in December 2010. (Possibly March – May 2011) F Ecoinformatics International Technical Collaboration is involved in planning and content development F A Technical Infrastructure Working Group will “focus on the technical components of the environmental information federation frameworks, addressing information and communications technology interoperability, connectivity, data standards, data format and content standards, and other such issues. This includes standards for the capture, description, and structuring of scientific data, and the development and delivery of various products and services. ” F
W 3 C, Web 2. 0, Web 3. 0 View F Suppose Sir TBL gave a presentation at the EOE Summit – Likely topic: u u Linked Data (aka Linked Open Data) – spirited inspirational presentation like he recently gave at TED and Gov 2. 0 conferences. Publishing Linked Data involves 3 basic steps n n n u 1. Assign URIs to the entities described by the data set and provide for dereferencing these URIs over the HTTP protocol into RDF representations. 2. Set RDF links to other data sources on the Web, so that clients can navigate the Web of Data as a whole by following RDF links. 3. Provide metadata about published data, so that clients can assess the quality of published data and choose between different means of access. Other Topics …
How to Do That … Suppose the United Nations Environmental Program wants to take such an approach to developing virtual state of the environment reports for countries (or other areas) around the world. F What technology infrastructure could help? F Essential components F u u u Data (preferably in RDF) Vocabulary, e. g. , ontology Metadata, e. g. , Vocabulary of Interlinked Datasets (voi. D)
Building Shared Vocabularies Moving from Limited Semantics to Fuller Semantics fuller semantics ty i iv s es s tic g a re c sin pr x e Formal Ontology OWL, Common Logic an m se Conceptual Model ER/UML In Concepts and relations expressed in an ontology representation language that provides formal semantics (i. e. , specifies logical inferences). Concepts and relations among them in a modeling language Taxonomy/Thesaurus Terms (possibly with definitions) & relations between terms Glossary Keywords Terms associated with definitions (concepts) Terms limited semantics
Connection between imagery and ontology Semiconductor Facility Pipe Water Treatment Facility g ss Buildin Proce Air Handling Facility ilding ocess Bu Pr Roof Warehouse Tank y ssagewa Pa uilding Enclosed Vehicle Parking/ tration B Process Building Storage dminis A Vehicle Parking/Storage Wall Flag pole Road Lawn le Vehic ing /Park rage Sto Fence Gas Utilities Facility Flare
Connection between imagery and ontology Does A=B? If True, imaged object is modeled entity. B E= route instrument U ha L s. VA Road Pipe Tank Flare Flag pole Flagpole Roof ing cess Build ing Pro Build Processti. Building istra on in Administration Building Warehouse rking a Warehouse/P building rage eway o cl lo St Storage/Parking Vehicle hince sed Passag Ve E covering Roof Fence Enclosed Passageway Wall barrier has. V AL land. UE = A cover/land use processing facilities Fence Lawn Wall Semiconductor Facility Air Handling Facility Water Treatment Semiconductor Facility Air Handling Facility Lawn Gas Water Treatment Facility Utilities Gas Utilities Facility
Ontology Collaboration Tool Capabilities and Services FRegister ontologies for use and development – upload & download FBrowse ontologies to find concepts, terms, definitions FVisualize ontology neighborhoods & hierarchy to root FDisplay metadata describing each ontology FCalculate metrics – classes, depth, definitions, structure, … FComment on concepts, relations, definitions, things to add, gaps… FMap between concepts in different ontologies FCreate views of ontologies that highlight important content FProvide an interface to query and download information, e. g. , SPARQL endpoint FLink ontology terms to resources (e. g. , simulations, algorithms, models)
Ontology Collaboration Tool Services Find and download selected ontologies F Answer questions about concepts/terms: F u u Find terms in one or more ontologies. What is the definition? What is the preferred term for a concept? What other concepts/terms are related? Find resources (data, simulations, algorithms, models) through ontology mediated search F Annotate text using ontology terms – direct, mapped, or through inference. F …. (movie) F
Browse Ontologies
Ontology Metadata
Ontology Metrics
Search for terms in Multiple Ontologies
View Concept Definition and Details
Visualize Ontology Neighborhood
Write Comments, Notes and Reviews … Subscribe to any Changes in oOntologies, Concepts, etc. Participants can use the “Notes” and “Review” functionality to: • Comment on classes, relations, definitions, gaps, … • Provide feedback to ontology authors • Reach consensus on ontology decisions • Review ontologies and their components • Subscribe to be notified of updates
Make Mappings between Ontologies
Projects Sign-up for Ontologies
Annotate Text with Ontology Terms Direct Annotations Extended annotations generated from the ontology is_a transitive closure.
Ontology Enabled Resource Discovery Data, models, simulations, …
Metadata for LOD “In order to support clients in choosing the most efficient way to access Web data for the specific task they have to perform, data publishers can provide additional technical metadata about their data set and its interlinkage relationships with other data sets …. The Vocabulary Of Interlinked Datasets … defines terms and best practices to categorize and provide statistical metainformation about data sets as well as the linksets connecting them. ” -- Tim Berners-Lee, Massachusetts Institute of Technology, et al F voi. D is a vocabulary and a set of instructions that enables the discovery and usage of linked datasets. A dataset is a collection of data, published and maintained by a single provider, available as RDF, and accessible, for example, through dereferenceable HTTP URIs or a SPARQL endpoint. Based on the voi. D vocabulary this document explains how to use voi. D in a practical setup, for both data consumers and data providers. -- from voi. D Guide
voi. D is Extensible The voi. D vocabulary is extensible. F It may be useful to extend it as needed for evaluating and documenting data for environmental decision making. F u E. g. , EPA data standards, ISO/IEC 11179 data descriptions
A Practical Example F Provided by Pasky Pascual, EPA u Inspired by his article: “Evidence-based decisions for the wiki world”, Pasky Pascual, International Journal of Metadata, Semantics and Ontologies (IJMSO) Volume 4 - Issue 4 – 2009 DOI: 10. 1504/IJMSO. 2009. 029232 u u He provided data for the Gulf of Maine LBNL is using this to demonstrate environmental linked data and voi. D files. n voi. Der software creates ISO/IEC 11179 type metadata on the dataset. This can be used to transform data into RDF and can be transcribed into voi. D descriptions.
Preserving the Comparability of Sensor Data A Possible Use Case Charles S. Spooner, US EPA ESIP 2010 Winter Meeting Washington, DC
Water Quality Data • Overwhelmingly an investment by public agencies • Our goal is to preserve that investment for future use recognizing that: – the value of good data increases over time – the value of undocumented data decreases quickly
Proposed Use Case • Compare metadata fields for 1. Existing WQ sensors in different settings • • • Vertical profilers Data Flow systems Autonomous Underwater Vehicles 2. The WQX Schema 3. Sensor Workgroup Data Elements – AQ, Calibration, Operator Competence 4. Water ML 5. Sensor ML
Use Case Results Source: Charles Spooner, ESIP Water Cluster Presentation, January 2010
Toxicity Data for MA, ME, NH
Fit into W 5 H Observations are Events Who: collecting agency What: observable measured When, Where How: method/protocol use rdf: value for measurement
voi. Der - Creation of ISO/IEC 11179 Metadata From Gulf of Main Toxicity Data Files (. xls) Results from a “clean” file:
voi. Der - creation of voi. D files From Gulf of Main Toxicity Data Files (. xls) Results from a “clean” file: Where When What Value
voi. Der - Creation of ISO/IEC 11179 Metadata From Gulf of Main Toxicity Data Files (. xls) Results from a “messy” file:
voi. Der - creation of voi. D files From Gulf of Main Toxicity Data Files (. xls) Shared ontology: @base <http: //xmdr. org/ont/toxicity. owl>. @prefix obs: <http: //xmdr. org/ont/observations. owl>. <> an owl: Ontology; owl: imports <http: //xmdr. org/ont/observations. owl>. : Toxicity a owl: Class; owl: sub. Class. Of obs: Observation. : species a owl: Object. Property; rdf: domain : Toxicity; rdf: range <http: //purl. bioontology. org/ontology/NCBI_NMO/Species>. . RDF data: @prefix tox: <http: //xmdr. org/ont/toxicity. owl>. . <> an owl: Ontology; owl: imports <http: //xmdr. org/ont/toxicity. owl>. <#Place-1> a geo: Point ; geo: lat 43. 1 ; geo: long -70. 77 ; rdf: label “Spinney Creek”. <#_2> a tox: Toxicity ; w 5 h: where <#Place-1> ; w 5 h: when “ 1985 -04 -16”^^xsd: Date ; tox: Species ncbi: Mytilus ; rdf: Value -58.
Practical Example: Ontology + Metadata in Use Sci. Scope F F F Sci. Scope shows the use of a water ontology, linked to water “variables” (data elements), with metadata that describes the data, and an easy to use geographic interface Demonstrates capabilities that help users to discover, evaluate, and access water data from millions of sensors for analysis, presentation, assessment, …. Shows use of metadata to describe the data Developed in collaborative effort between Microsoft Research, Berkeley Water Center (UCB), Lawrence Berkeley National Laboratory Use it at: Sci. Scope. org (Hosted by LBNL) See a movie: Sci. Scope_Movie. wmv
Sci. Scope – Ontology & Metadata put millions of sensors at your fingertips Metadata descriptions Browse geographical features from eco-regions and hydrology to geology Sci. Scope facilitates data discovery from 9. 5 million sensors in the USA operated by agencies such as USGS, EPA and NOAA offering observation results from late 1800’s to the current day Adapted from source: Bora Beran, Microsoft Research Ontology mediated discovery Find and retrieve historic and near real-time data about the environment from multiple databases Assemble data from multiple heterogeneous databases with ease
What is behind Sci. Scope? Knowledge Base F Relationships are stored as RDF triples in a relational database F ‘Escherichia coli’ = ‘E. coli’ is-a ‘Indicator Organism’ ‘Nitrogen’ is-a ‘Macronutrient’ is-a ‘Nutrient’ ‘Hypoxia’ is. Measured. Using ‘Dissolved. Oxygen’ ‘Hypoxia’ is. Related. To ‘Eutrophication’ Supports transitive, symmetric and inverse properties F Inferred statements are pre-computed F Source: Bora Beran, Microsoft Research
Inference In Sci. Scope Transitive ‘Nitrogen’ is-a ‘Macronutrient’ is-a ‘Nutrient’ Inference: ‘Nitrogen’ is-a ‘Nutrient’ F Symmetric ‘Hypoxia’ is. Related. To ‘Eutrophication’ Inference: ‘Eutrophication’ is. Related. To ‘Hypoxia’ F Inverse ‘Macronutrient’ is-a ‘Nutrient’ Inference: ‘Nutrient’ is. Broader. Than ‘Macronutrient’ F Source: Bora Beran
What is Behind Sci. Scope – Linking Concept Systems to Data
What is behind Sci. Scope? Geographical Features Catalog F Collection of features such as dams, aquifers, geologic formations, watersheds, sensors F Based on data and maps from USGS, EPA, National Atlas Source: Bora Beran
Show Sci. Scope Movie
Acknowledgements F F F Bora Beran, Micro. Soft Research Kevin Keck, LBNL Glenn May, LLNL Mark Musen, Natasha Noy, et al, Stanford Pasky Pascual, EPA Charles Spooner, EPA This material is based upon work supported by the National Science Foundation, under Grant No. 0637122, by USEPA and by DOE. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation, DOE, or USEPA.
b08eeab04d6c30c56f53d69fd741580b.ppt