Tuesday, January 5, 2010

Semantic Web

An updated SVG of the FOAF logo. I created the...Image via Wikipedia

The semantic web promises to define the meaning of its content (semantics) making computer programs able to perform all the tedious work of searching, comparing and combining bits and pieces of information - all the tasks we have to do by ourselves in order to find answers to complex questions.

Attempts to make this possible includes languages: Resource Description Framework (RDF), Web Ontology Language (OWL), and Extensible Markup Language (XML); data interchange formats (e.g. RDF/XML, N3, Turtle, N-Triples), and notations such as RDF Schema (RDFS). Some believe that even simpler unambiguous formats will be part of the semantic web.

An early example of a semantic web application, FOAF (an acronym of Friend of a friend) expressed using RDF and OWL is a descriptive vocabulary, an ontology describing people, their activities, and relations to each other. Check, for example, semantictweet.com that allows to turn Twitter accounts into FOAF profiles.

Simpler components of semantic web include meta-tags - an extension of metadata in the form of tags used to describe Web pages' content in the early days of web design. This could be labels with keywords, name of the page author, description or metadata representing sets of facts. Ontologies are often used to generate metadata and mapping between vocabularies.
Zemanta system semantically filters content of the page to automatically tag it or suggest other relevant content. Semantic standards like Common Tag - developed jointly by Zemanta, Metaweb, and Yahoo! - add semantic meaning to tags expressed using RDFa, making Web content more discoverable and decentralized. With Common Tag, content is tagged with unique, well-defined concepts - everything jaguar the animal is tagged with one concept for jaguar the animal vs. the car.

Ontologies defined as "an explicit specification of a conceptualization" are about the kind of things existing in the domain and their relations to each other. There could be no perfect scheme, however. In working classification systems, success is not "Did we get the ideal arrangement?" but rather "How close did we come, and on what measures?" There are no fixed shelves and perfect arrangements. It would not hurt to rearrange things from time to time, add a few secondary links and subdirectories. "This book is mainly about the Balkans, but it's also about art, or it's mainly about art, but it's also about the Balkans." The classification can not be perfect - except very small and specialized domains, stable entities, clear edges, authoritative sources and a lot of money to support the authoritarian system. This will never work for the Web - with many uncoordinated users that don't think alike and are naive classifiers.

Web users have very different tagging strategies too - but all together they are creating value for one another, by connecting similar tags.

Some of most often mentioned examples of semantic web application include Freebase - an open database where anyone can perform complicated queries. It is like a structured version of Wikipedia, combining the advantages of free text and relational databases. It's qualified users can not only add the content but also change the structure itself modifying the definitions of existing types. With many qualified users both the content and structure will evolve over time.

Even though some sources claim that Freebase is built around an exhaustive ontology, it's more like a collection of ontologies or rather tags. Unlike the W3C approach to the semantic web, which starts with controlled ontologies, Metaweb adopts a folksonomy approach, in which people can add new categories (much like tags), in a messy sprawl of potentially overlapping assertions.

See more on the State of Linked Data in 2010:
From the Semantic Web to the Web of Data: ten years of linking up

An example of a semantic data storage tool able to analyze networks and events in large volumes of structured and unstructured data is AllegroGraph. this RDF graph database can be, for example, used for GeoSpatial and Temporal Reasoning performing complicated queries against news articles scraped from Google

NYCSW GeoSpatial, Temporal Reasoning with AllegroGraph from Morton Swimmer on Vimeo.

Reblog this post [with Zemanta]

No comments :

Post a Comment