IBM 1620 data processing machine, 1962
Who is this?
The Web
The Web accompanies the transition from an industrial to an information society and provides the infrastructure for a new quality of information handling regarding acquisition as well as provisioning
- high availability
- high relevance
- low cost
The Web penetrates society
- Social contacts (social networking platforms, blogging, ...)
- Economics (buying, selling, advertising, ...)
- Administration (eGovernment)
- Work life (information gathering and sharing)
- Recreation (games, role play, creativity, ...)
- Education (eLearning, Web as information system, ...)
The current Web
Immensely successful.
- Huge amounts of information and data.
- Syntax standards for transfer of structured data.
- Machine-processable, human-readable documents.
BUT:
- Content/knowledge cannot be accessed by machines.
- Meaning (semantics) of transferred data is not accessible.
Limitations of the Web
Too much information with too little structure and made for human consumption- Content search is very simplistic
- → future requires better methods
- in terms of content
- in terms of structure
- in terms of character encoding
-
→ future requires intelligent information integration
- → requires automated reasoning techniques
What Google does not find
There are many information needs current search engines can not satisfy:
- Apartments for rent close to well rated Thai restaurants
- Bi-lingual English-German child care in Berlin reachable in 15 minutes from my place of work
- Kid-friendly holiday destinations with culture and sports activities
- Researchers working in south-east Asia on information retrieval topics
- ERP service providers with offices in Vienna and Berlin
- ...
We have subconsciously learned not to ask search engines such questions.
In principle, all the required knowledge is on the Web – most of it even in machine-readable form. However, without automated data integration, processing (and reasoning) we cannot obtain a useful answer.
What's the problem with the Web
- inability to integrate and fuse information from different sources
- there is lack of comprehensive background knowledge to interpret information found on the Web
- current Web search is restricted to text in a certain language - there are many “smaller” languages with much less information available than in English
Basic ingredients for the Semantic Web
- Open Standards for describing information on the Web
-
Methods for obtaining further information from such descriptions
We’ll talk about these matters in this course.
Data Models, Access & Integration
| Data Integration | Enterprise Information Integration sets of heterogeneous data sources appear as a single, homogeneous data source |
Data Warehousing
|
Research
|
Data Web
|
| Data Access | Object relational mappings (ORM)
|
Procedural APIs
|
Query Languages
|
Linked Data
|
| Data Models | RDBMS
|
LOD Cloud May 2007
LOD = Linked Open Data
LOD Cloud October 2007
LOD Cloud February 2008
LOD Cloud September 2008
LOD Cloud March 2009
Colours indicate different domains; e.g.: orange = social networks
LOD Cloud September 2010
LOD Cloud September 2011
LOD Cloud August 2014
First update after 3 years – what do you think were the reasons?
LOD Cloud February 2017
The Web of Data
- >70 bilion facts
- covering many different domains (life-sciences, geo, user generated content, government, bibiographic, ...)
Map to the Semantic Web
The Semantic Data Web Stack
… also known as “layer cake”
How Mark A. Greenwood realised the Semantic Web:
URIs and Unicode
- URI = Uniform Resource Identifier
- Used to create globally unique names for resources
-
Every object with clear identity can be a resource
- Books, places, organizations ...
- In the books domain the ISBN serves the same purpose
- IRIs: Unicode-aware extension of URIs (I = Internationalized)
Resource Description Framework – RDF
Information is represented in RDF in triples (also called statements, facts):
- Modeled on linguistic categories, but not always consistent
- Allowed assignments:
- Subject: URI or blank node
- Predicate: URI (a.k.a. property)
- Object: URI, blank node or literal
- Node and edge labels should be unambiguous, so that the original graph is reconstructable from a list of triples
RDF Schema
Not all triples make sense:
Cinema AlbertEinstein 2012
How can we constrain the use of RDF?
RDFS (S = “Schema”) allows to define classes, properties and restrict their use.
SPARQL – Query Language for RDF
SELECT * WHERE { jwebsp:John foaf:knows ?friend }
Web Ontology Language – OWL
- OWL: acronym for Web Ontology Language, more easily pronounced than WOL
- family of languages for authoring ontologies
- since 2004, OWL 2.0 since 2009
- Semantic fragment of FOL
Features
- Instantiation of classes by individuals
- Concept hierarchies (taxonomies, inheritance): classes, terms
- Binary relations between individuals: Properties, Roles
- Properties of relations (e.g., range, transitive)
- Data types (e.g. Numbers): concrete domains
- Logical means expression
- Clear semantics!
RDFa Content Editor – RDFaCE
supports the automatic semantic annotation of texts
Literature
- Pascal Hitzler, Markus Krötzsch, Sebastian Rudolph: Foundations of Semantic Web Technologies, Chapman & Hall/CRC, 2009, 455 pages, hardcover, ISBN: 9781420090505, http://www.semantic-web-book.org
- Amit Sheth, Krishnaprasad Thirunarayan: Semantics Empowered Web 3.0: Managing Enterprise, Social, Sensor, and Cloud-based Data and Services for Advanced Applications (Synthesis Lectures on Data Management), Morgan & Claypool Publishers (December 19, 2012), ISBN: 1608457168
- Tom Heath, Christian Bizer: Linked Data (Synthesis Lectures on the Semantic Web: Theory and Technology), Morgan & Claypool Publishers; 1 edition (February 20, 2011), ISBN: 1608454304. http://linkeddatabook.com
Questions
All the corresponding questions for the Introductions are covered in the Questions part of the Deck.