9. Mika - semantic technology - Research Information Network

hurtpotatocreekSecurity

Nov 5, 2013 (3 years and 8 months ago)

60 views

Semantic Technology

supporting science

Peter Mika / Dept. of Computer Science / Vrije Universiteit, Amsterdam

Overview

Two systems



One technology


Many possibilities

flink

networks in science

Flink

Social network data collection, aggregation,
storage and visualization

Target: the Semantic Web community

Semantic Web technology

Ontology
-
based representation and reasoning

Try it

http://flink.semanticweb.org

Open source

(in part
)

Elmo API for Sesame

http://www.openrdf.org

FOAF profiles


Web

Emails

Publications

Representation,

storage and
reasoning

Presentation
and Analysis

Data acquisition

Sesame

Sesame

Sesame

Sesame

Browsing

Subcommunities

Associations between
research topics

Geographic visualization

Network analysis

Network measures vs.

status vs. performance

SWWS, ISWC chair (4)

W3C co
-
chairs (2)

Journal of Web Semantics (4)

IEEE Intelligent Systems (3)

openacademia

metadata for the masses

Metadata micro
-
management

Repository for small research groups

Software you can download and install

Distributed system (unlike CiteSeer, DBLP…)

Open source


As easy as…

Maintaining a BibTex/Endnote file for yourself

Optionally: filling out a form to create a personal/group profile


Instant gratification

For the researcher: publication list
and RSS feed
for homepage

by
adding a
<LINK> tag and one

line of JavaScript

For the group: reporting, dissemination on group homepage etc.


http://openacademia.org



Pimp your homepage

Query interface

Applying the BibTex stylesheet

We got tagclouds!

And social networks.

RSS feeds, live bookmarks

Publication list for homepage

Architecture

Can be another
openacademia
s
erver
!

Can be remote
server
!

XSLT transformation
to produce to
produce HTML,
BibTex etc.

Semantic

technologies

The benefits:

modelling & aggregation

Explicit

RDF/OWL allows to express and reason with what it means
for two things to be the same (smushing)

Extendible

Designed to be distributed both in terms of schema and data

Mappings between different schemas can also be expressed
in the language

Flexible

Mappings can be partial, robustness*

Standard

Standard languages (RDFS, OWL, SPARQL)

Standard vocabularies (DC, PRISM, SWRC)

Standard protocols (SPARQL)

The drawbacks

Limited expressivity

e.g. complex inverse functional properties

e.g. swrc:page, prism:startingPage and prism:endingPage

Ontology
-
based interchange is still partly social
engineering

Scalability



What about Web 2.0?

Folksonomies are ontologies

Large number of
individual tagging actions
result in the emergence
of the semantics of tags


Lightweight, dynamic*
ontologies





P. Mika.
Ontologies are us: A unified model of
social networks and semantics
. In: Proceedings of
the Fourth International Semantic Web Conference
(ISWC 2005), Yolanda Gil, Enrico Motta, Richard V.
Benjamins and M. A. Musen (eds.) , Lecture Notes in
Computer Science no. 3729, page 122
-
136, Galway,
Ireland, November, 2005

Tagging

Tagging interests in flink, topics of publications in openacademia
(also Connotea, CiteULike, bibsonomy etc.)


Tag interchange is problematic in general



flickr:ajax = del.icio.us:ajax ?



flick:ball:Peter = flick:ball:John ?



flick:ball:Peter:1990 = flick:ball:Peter:2006 ?


More flexible than controlled vocabularies

Tracks the evolution of the language better


Should work for scientific objects (publications, presentations
etc.)

Users have the same object in mind when tagging, limited
community (scientific jargon)



Blogging, semantic wikis

openacademia imports comments about
publications

Required: blog search (auto
-
discovery?)

Semantic wikis are promising

Metadata directly in RDF

Syntactic metadata for now (who
commented on what, what time)


Example: Bibster

P2P bibliography sharing system

Each peer has an RDF triple store with
publication metadata

Advanced query routing based on semantic
models of the content and user interests

Outcome of the EU IST project SWAP and
winner of a number of awards, featured
on Slashdot

No one uses it.

Software you install


and keep running



openacademia p2p

Servers of research groups are
networked

Web
-
based infrastructure


VUA

Stanford

openacademia.org/

servers.rdf

p2p spirit

flink and openacademia can be ‘edited’
by anyone

1.
Create descriptions of publications,
personal metadata, group and event
definitions

2.
Let our crawler find it

future

bright

Trends

Changing form of publishing

Demise of the journal as distribution channel

Community reviewing

Demise of the journal as quality seal

The semantic conference

e.g. ESWC 2006


In general:


More and more data


Increased connectedness of data sources


Productivity




Do you still go
to the library?

Do you still
read journals?

Online repositories

Researchers

Who is involved?

Publishers

Libraries