Semantic Web concepts
Edinburgh
http://www.w3.org/2005/Talks/0517-boit-tbl/
Tim Berners-Lee
MIT Computer Science & Artificial Intelligence Laboratory (CSAIL)
Director, World Wide Web Consortium
Three 'hot' directions of Web development
- Mobile Web
- Web Services
- Semantic Web
What is Semantic Web?
- Data integration across application, organizational boundaries
How does it work?
- Apply power of URIs to concepts of relational data
- Model real things, not documents or database tables
SW: Everything has a URI
Don't say "colour" say
<http://example.com/2002/std6#col>
The relational database
![A database's row, column and cell are subject, property and value.](../../../DesignIssues/diagrams/spv-table.png)
The element of the Semantic Web
![arrow tail, body and head are l are subject, property and value.](../../../DesignIssues/diagrams/spv-arrow.png)
- Can be encoded in XML
- Simplicity and mathematical consistency
- This is called Resource Description Framework (RDF)
Semantic web includes tables,...
![Arrows can make a table, an arrow from each row to each value](../../../DesignIssues/diagrams/arrow-table.png)
...trees
![Arrows can make a table, an arrow from each row to each value](../../../DesignIssues/diagrams/tree.png)
... everything
![Arrows can make a table, an arrow from each row to each value](../../../DesignIssues/diagrams/tree-and-table2.png)
Applications connected by concepts
![Its like a metro, the way the lines of common concepts connect the stations of different applications](../../../DesignIssues/diagrams/SemWebAppMetro.png)
Fractal Web of concepts
- Across boundaries of scale -- personal, group, global
- Varying access levels
- Tension between local and global standards
- Society is a fractal tangle, so must SW be.
- Personal interactions on multiple scales
Example: Drug discovery
![FOAFm OMM, UMLS, SNP, Uniprot, Bipax, Patents all have some overlap with drug target ontology](../../../DesignIssues/diagrams/lifesci/endq/img8.gif)
2003: Enterprise Application Integration
![N by N problem without RDF](../../../DesignIssues/diagrams/nbyn.png)
RDF: Enterprise Integration hub
![N by 1 problem with RDF](../../../DesignIssues/diagrams/nby1-no-bus.png)
Global Integration bus
![N by 1 problem with RDF](../../../DesignIssues/diagrams/nby1.png)
Clients of the RDF bus
New data applications can be built on top of RDF bus, for example:
![db to sw](../../../DesignIssues/diagrams/sw-clients.png)
Components: Adapting random files
Keep your existing systems running - adapt them
![db to sw](../../../DesignIssues/diagrams/sw-adapt-etc.png)
Components: Triple store
Virtual severs actually figure stuff out as well as look up data
![db to sw](../../../DesignIssues/diagrams/sw-store.png)
Adapting SQL Databases
Keep your existing systems running - adapt them
![db to sw](../../../DesignIssues/diagrams/sw-db-to-sw.png)
Adapting XML
Remeber- RDF on an HTTP server can always be virtual
![db to sw](../../../DesignIssues/diagrams/sw-adapt-xml.png)
Adapting XML: GRDDL
Remeber- RDF on an HTTP server can always be virtual
![db to sw](../../../DesignIssues/diagrams/sw-grddl.png)
Components: Smart servers
Virtual severs actually figure stuff out as well as look up data
![db to sw](../../../DesignIssues/diagrams/sw-infer.png)
Infrastructure roadmap
Stack of expressive power
![architectural layers](../../../DesignIssues/diagrams/sw-stack-2005.png)
Challenges: Deployment
- Scientific data:
- Integration across fields
- Eg Life sciences, drug discovery; Geospatial
- Pointers to experimental conditions, sources, algorithms
- Electronic Commerce:
- Well-defined meaning for documents
- Catalogs, prices, rates, specifications as data
- Personal Information Management: Calendar, Photos, Money, etc
- Common platform for database, inference, etc etc
- Index rules & ontologies if not data: new algorithms
- New challenges on top of Semantic Web - AI
Snowball problem: Techniques
Making communities
- Build common ontologies
- Adapt legacy systems - GRDDL etc.
- Accumulate & publish Best Practices
- Showcase early adopters
- Motivate by vision
- Motivate by other, short term, return on investment
Challenge: Logic languages
- Integration of multiple logics on same
- Quotation & self-references makes FOL tricky
- Statements about statements
- Exchange of proofs between dissimilar inference engines
Challenges: Looking ahead
Semantic-Web based research
- What about a web of rules? Web of logic?
- Indexing that, inference using indexes.
- Delegated query
- Ultimate User Interface to the Semantic Web
- Speech and Graphic UI info hints in ontologies
Non-monotonic/Fuzzy things
Not Semantic Web phase one.
- Extraction of semantics from exiting natural language
- Probabilistic heuristics for ontology matching
Deployment: Flagship fields
- Scientific data:
- Integration across fields
- Eg Life sciences, drug discovery; Geospatial
- Pointers to experimental conditions, sources, algorithms
- Electronic Commerce:
- Well-defined meaning for documents
- Catalogs, prices, rates, specifications as data
- Personal Information Management: Calendar, Photos, Money, etc
Deployment: Techniques
Making communities
- Build common ontologies
- Adapt legacy systems - GRDDL etc.
- Accumulate & publish Best Practices
- Showcase early adopters
- Motivate by vision
- Motivate by other, short term, return on investment
Deployment: Action
IT Management
- Model the company business
- Use existing ontologies where they meet your needs
- Develop your own ontologies
- Work with other groups, other companies
- Map data sources to RDF and SPARQL: databases, XML, spreadsheets,
etc...
- Deploy visualizers, inference, validation etc tools
- Don't change the way you do things underneath
- Be aware of SW deployment curve
World Wide Web Consortium
Global organization for a global infrastructure
- Users, Vendors, Academics meet
- Pool investment of energy
- Forming consensus: hard work but invaluable
- Well-honed process and patent policy
- Decentralized architecture
- Working on SPARQL, best practice
- Starting work on Rules, life sciences