MultimediaN demonstrator: culture in context

Why do I need the Semantic Web in the first place?

So you have heard all about XML, RDF, RDFS, OWL, WordNet in RDF, Ontologies, Description Logics, Rule Languages ...

But what is it all good for in practice?

My two cents:

Interoperability: data interchange and data integration
Pushing part of the reasoning from the application layer down the stack to the middleware layer

Partners: ICN, DEN, UvA, VU, CWI
Improve "in context" search and presentation of search results by
- Linking multiple collections (often standalone)
- Reusing existing background knowledge (rich domain!)
Current status:
- Collections: Rijksmuseum A'dam (750), Artchive.com (3000), Rijksmuseum Volkenkunde (250.000), ...
- Thesauri: Getty AAT, TGN, ULAN; SVCN, WordNet, ...
- In total about 10 million RDF statements

KLIMT, Gustav
Portrait of Adele Bloch-Bauer I
1907
Oil and gold on canvas
138 x 138 cm
Austrian Gallery, Vienna

MUNCH, Edvard
The Scream
1893
Oil, tempera and pastel on cardboard
91 x 73.5 cm
National Gallery, Oslo

Open Web standards
- XML, XHTML, SVG, RDF/OWL, SPARQL
SWI-Prolog’s semantic web package
- SPARQL interface, RDF/OWL partial reasoning
AJAX for rich Web client interaction
- Semantic keyword suggestion
- Dynamic tree browser for structured taxonomies
Java applets for other interaction types
- Picture point: annotating image content

Get the data!
- Musea are not used to share their data
- Thesauri take sometimes (100s of) years to build: do we just give that away?
- Actual data often maintained by other companies
Get the data in RDF and/or OWL
- Covert both museum collections and thesauri to RDF
- Guidelines? (SKOS, work of Van Assem et al., W3C SWBPD WG)
- Minting URIs (you need a lot of them!)
  - for somebody else's concepts
  - for non-informational resources (httpRange-14, slash vs hash, t-d-b.org, Connoly's work, ... )
  - Example collection property for Artchive virtual collection ...
Link & enrich the data

Parse annotation to find matches with thesauri terms
- E.g. match artists to ULAN individuals
Artists-style links
- AAT contains styles; ULAN contains artists, but there is no link
  - Learn link from corpora
  - Derive it from other annotations
- Domain-specific rules/reasoning needed
  - see example in SWRL doc
  - Painters may have painted in multiple styles
Annotation scheme: annotations are not "just" instances of your ontology (DC, VRA, EC)
Image annotations (PicturePoint, MMSEM XG)

Choosing the right model theory
- RDF, RDFS, OWL Lite, DL, Full
- data vs processing
- Reference vs application ontologies

User testing: domain experts & general public
More collections & thesauri
- We start getting them from other projects, others start using ours!
New search paradigms (relation search, space/time reasoning)

Acknowledgments: Guus, Mark and many others

The Semantic Web Artificial Intelligence on the Web
It relies on centrally controlled ontologies for �meaning�
- as opposed to a democratic, bottom�up control of terms
One has to add metadata to all Web pages, convert all relational databases, and XML data to use the Semantic Web
It is just an ugly application of XML
One has to learn formal logic, knowledge representation techniques, description logic, etc
It is, essentially, an academic project, of no interest for industry
�

The Semantic Web Artificial Intelligence on the Web
It relies on centrally controlled ontologies for �meaning�
- as opposed to a democratic, bottom�up control of terms
One has to add metadata to all Web pages, convert all relational databases, and XML data to use the Semantic Web
It is just an ugly application of XML
One has to learn formal logic, knowledge representation techniques, description logic, etc
It is, essentially, an academic project, of no interest for industry
�