semantic search for enterprise 2.0

14
Copyright 2008 Digital Enterprise Research Institute. All rights reserved. Digital Enterprise Research Institute www.deri.ie Semantic Search for Enterprise 2.0 Alexandre Passant 1 , Philippe Laublet 2 , John Breslin 1 , Stefan Decker 1 1 Digital Enterprise Research Institute, NUI Galway 2 LaLIC, Université Paris-Sorbonne, France SemSearch09, WWW09, Madrid 21th April 2009

Upload: alexandre-passant

Post on 05-Jul-2015

1.883 views

Category:

Technology


2 download

DESCRIPTION

SemSearch09 workshop at WWW2009, April 21th 2009- http://km.aifb.uni-karlsruhe.de/ws/semsearch09/ - Paper available at: http://km.aifb.uni-karlsruhe.de/ws/semsearch09/semse2009_25.pdf

TRANSCRIPT

Page 1: Semantic Search for Enterprise 2.0

Chapter ♥ Copyright 2008 Digital Enterprise Research Institute. All rights reserved.

Digital Enterprise Research Institute www.deri.ie

Semantic Search for Enterprise 2.0

Alexandre Passant1, Philippe Laublet2, John Breslin1, Stefan Decker1

1 Digital Enterprise Research Institute, NUI Galway 2 LaLIC, Université Paris-Sorbonne, France

SemSearch09, WWW09, Madrid

21th April 2009

Page 2: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Enterprise 2.0

  Social media in a corporate context   “The use of emergent social software platforms within

companies, or between companies and their partners or customers”

  The SLATES paradigm   Search

  Links

  Authoring

  Tagging

  Extension

  Signals

Page 3: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Main issues

  Enterprise 2.0 can be used to foster collaboration and social intelligence, but raises various issues

  Information fragmentation   Description of a project on a wiki, minutes of meetings on

blog posts, information about partners on RSS feeds, etc .

  The gap between documents and data   Valuable information in wikis, but hard to efficiently get it,

e.g “list all companies involved in project X since 2008”

  Tagging issues   Ambiguity, heterogeneity, lack of organization and gap of

tagging behaviors depending on expertise

Page 4: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Proposed solution

  Considering Enterprise 2.0 at the level of semantics   The SemSLATES approach: middleware for Enterprise 2.0

  Following the RDF bus approach   Add-ons for existing applications

  RDF(S)/OWL and SPARQL

  User-interfaces and applications

on the top of it

Page 5: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Use-case

  EDF R&D: Blogs, wikis, RSS feeds   Extensions for data integration, enabling semantic mash-

ups and semantic search

  Common semantics for various applications   SIOC and related vocabularies to model the social

interactions within the user communities

  From documents to structured and interlinked data   Lightweight ontologies (SKOS, FOAF extensions …)

  Extending the Wiki platform to a Semantic Wiki system

  Tagging issues   Semantic tagging with MOAT, i.e. “Tag with URIs”

Page 6: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Exposing SIOC data

  Enable common semantics for user-generated content   From various applications, completely automated

Page 7: Semantic Search for Enterprise 2.0

Inline macro

Simple autocomplete

field

Complex instance field

Digital Enterprise Research Institute www.deri.ie

From documents to RDF data

  UfoWiki   Wiki interface including forms mapped to ontologies for

collaborative instances management

  Live SPARQL-autocompletion to reuse URIs between wikis

Page 8: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Semantic tagging

  MOAT - Meaning Of A Tag   A lightweight model and framework to bridge the gap

between tagging and semantic indexing

  Tags mapped to ontology instances created from the Wikis (via their label)

  User-interface for validation / disambiguation (if needed)

  Ability to link new tags to existing instances

Page 9: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

A complete interlinked graph

Wiki page 2Blog Post 1

hyperlink

a:EDF

f:Company

g:3017382/

g:Featurea:Energy

rdf:type rdf:typea:produces

g:locatedIn

:bpost_1

moat:topic

:wikipage_2sioc:links_to

:wiki_A

sioc:container_of

:alex

sioc:has_creator

creates contains

Wiki A

Sem

antic W

eb layer

Ente

rprise 2

.0 s

erv

ices

Social interactionswth SIOC

Semantic tagging with MOAT

Ontology populationwith semantic wikis

Page 10: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Enabling search

  How to integrate and query all this data ?   17000 instances of sioc:Post linked to 300 domain

ontology instances, on various applications

  A ping-based architecture with a central RDF store   Each component pings the store when creating /

updating / deleting RDF data

  REST-ful interactions using SPARQL / SPARUL

Semantic Middleware

RDF Store

SPARU

L

inte

rface

SPARQ

Lin

terf

ace

Ping system to store

new or updated data

External services using

stored data

Page 11: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Search interface

  End-user interface   Identifying relevant instance and retrieving information

from various sources, as well as related entities

  Hiding RDF(S)/OWL and SPARQL to the users

Page 12: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Enabling Semantic mash-ups

  Re-using RDF data from the LOD cloud internally   Low-cost Semantic mash-ups

  E.g. Geolocation of wiki instances thanks to Geonames

Page 13: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Conclusion

  Enterprise 2.0 enables social interactions and ease content-generation   But introduces new issues / emphasizes existing ones

  Semantic integration can help   Without having to rebuild the information system

  Lightweight add-ons, transparency for end-users

  Compared to existing information integration approaches   Use lightweight semantics (FOAF, SIOC, SKOS …)

  Consider the social aspect of Enterprise 2.0 both when creating and using RDF data

Page 14: Semantic Search for Enterprise 2.0

Digital Enterprise Research Institute www.deri.ie

Thank you !

  Any questions ?

  Contact   http://apassant.net

[email protected]