Upload
freddy-limpens
View
906
Download
1
Tags:
Embed Size (px)
DESCRIPTION
Citation preview
From folksonomies to structured knowledge representations: bridging Collaborative Web and Semantic Web
Des folksonomies aux représentations structurées de connaissances: faire le pont entre Web Collaboratif et Web
Sémantique
1
A D B S / C o o p é r a - o n & D é v e l o p p em e n t A t e l i e r We b S ém a n - q u e e t D é v e l o p p em e n t D u r a b l e – 3 1 . 0 1 . 2 0 1 1
Freddy Limpens fdy@pl-‐area.net
h0p://pl-‐area.net
2
From social tagging to folksonomies
Tags freely associated to resources …
… collected and shared on the web
3
… resul=ng in
FOLKSONOMIES
A mass of users for a mass of resources
Limita-ons of folksonomies
4
Spelling varia-ons of tags:
newyork = new_york = nyc
Lack of seman-c links between tags
Limita-ons of folksonomies
5
Lack of interoperability between social data repositories
Limita-ons of folksonomies
6
7
How to turn folksonomies ...
?... into
topic structures (thesaurus) ?
pollution
Soil pollutions
has narrower
pollutant Energy
related related
1. State of the art
8
9
State of the art
Involving users in tags structuring:
• Simple syntax to structure tags (Huyn-‐Kim Bang et al. 2008)
• Crowdsourcing strategy to validate tag-‐concepts mapping (Lin et al. 2010)
pollution
Soil pollutions
has narrower
pollutant Energy
related related
10
State of the art
Automa-c extrac-on of tag seman-cs:
pollution
Soil pollutions
has narrower
pollutant Energy
related related
11
Tags and Seman-c Web models
TAGS + SCOT + SIOC + FOAF for tags and tagging :
tags:Tagging #11111
sioc:Item h0p://www.windenergy.com
tags:taggedResource
scot:Tag #wind-‐energy
tags:associatedTag
foaf:Agent #freddy.limpens
tags:taggedBy
12
Tags and Seman-c Web models
Tagging = linking a resource with a sign
What is a tagging ?
"nature"!
picture shows "nature" (1) (2) (3)
place located l:england
edi=ng makes me : )
13
Tags and Seman-c Web models
NiceTag (Monnin et al, 2010):
Tagging as named graphs*
*Carrol et al. (2005)
nt:TaggedResource h0p://www.windenergy.com
nt:ManualTagAc=on (named graph)
nt:isAbout scot:Tag #wind-‐energy
sioc:UserAccount freddy
sioc:has_creator
sioc:Container delicious.com
sioc:has_container
14
Tags and Seman-c Web models
2 complementary seman=c enrichment:
wind-‐energy
renewable energy
windenergy
wind turbine
has broader
close match
has narrower
environment
related
Structuring tags as in a thesaurus (SKOS)
2. 1st Applica-on case : Corpus management at ADEME
15
16
Ademe scenario
Experts produce docs
+ tag Archivists
centralize + tag
Public audience read + tag
ADDING TAGS
Automatic processing
User-centric structuring
Detect conflicts
Global structuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment life-‐cycle
17
ADDING TAGS
Automatic processing
User-centric structuring
Detect conflicts
Global structuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment life-‐cycle
18
19
1. String-‐based metrics
pollution Soil pollutions
pollutant pollution
=> « pollution » related to « pollutant »
=> « pollution » broader than « soil pollutions »
renewable energy wind-‐energy
Alex
Delphine
Claire
Monique
Anne
⇒ Hyponym rela=ons (broader/narrower):
« renewable energy » broader than « wind-‐energy »
3. User-‐based associa-on
20
3. User-‐based associa-on
21
22
!"#$%&'"()&$!"#$*"&&'+)&$!"#$#,)--.*/$0"&."*1$!"#$&)-"1)($
,'+)&)($%2$/),!.3'&/$
Computed rela.ons are not always accurate
ADDING TAGS
Automatic processing
User-centric structuring
Detect conflicts
Global structuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment life-‐cycle
23
24
Capturing users's contribu-ons
Embedding structuring tasks within everyday ac.vity (searching e.g)
25
Capturing users's contribu-ons
ADDING TAGS
Automatic processing
User-centric structuring
Detect conflicts
Global structuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment life-‐cycle
26
27
Conflict detec-on
environment pollu=on narrower
John
hasApproved
Anne
hasApproved
broader
Monique
hasApproved
Delphine
hasApproved
Experimenta-on at ADEME
!"#$%&'#()*+,)
-../"012)34,)
516787691):;,)
<1=1&812):+,)
!"#$%&'("&$)*+,&-$'.$/'012/-$+'&3204$
28
ADDING TAGS
Automatic processing
User-centric structuring
Detect conflicts
Flat folksonomy
Structured folksonomy
Folksonomy enrichment life-‐cycle
29
Global structuring
30
Global map
Includes all points of view, highlights conflicts + consensuses
Referent choices
31 Choices of the referent user (archivists at Ademe e.g.)
ADDING TAGS
Automatic processing
User-centric structuring
Detect conflicts
Global structuring
Flat folksonomy
Structured folksonomy
Folksonomy enrichment life-‐cycle
32
Each point of view corresponds to a layer
33
Enriching individual points of view
Integra=ng others' contribu=ons: 1. Current user -‐> "Anne" 2. ReferentUser (e.g. archivists) 3. ConflictSolver (sohware agent) 4. Other individual users 5. Automatons (metrics)
BROADER
NARROWER
RELATED
CLOSE MATCH
environnement Search:
preoccupa=on environnementales
grenelle de l environnement
competences environnementales
environment
environmental
domaines environnementaux
Anne is looking for resources tagged "environnement"
34
2. 2nd Applica-on case : Leveraging the reuse of 2nd hand objects
35
How will I get rid of all these rusty coffee makers??
Paris
Nice
Digitazing the stock of 2nd hand shops
37
? ?
?
The goal : Finding semantically
Related tags To enhance searching
coffee maker
Seman-cally Enhanced catalog
The idea : Mapping tags With ontologies’ concepts
Hot liquid container
coffee maker coffee pot
tea pot
= subClassOf
Seman-cally Enhanced catalog
Results for "coffee maker":
coffee maker
Related results :
Results for "tea pot":
Results for "coffee pot":
1. The user enter "coffee maker"
Seman-cally Enhanced catalog
2. The system suggests addi=onal results thanks to seman=c rela=ons
5. Conclusion
41
42
What we do :
Help online communi=es
structure their tags wind-‐energy
renewable energy
sustainability
wind turbine
has broader
related
has narrower
environment
related
An approach to bridge tagging with Seman-c Web:
Automa-c processing of tags:
User interface to capture tag structuring embedded in every-‐day tasks
Implementa-on within ISICIL solu=on (tagging server)
43
Our contribu-ons:
• 1st scenario:
• More user interfaces
• test within ISICIL (ANR) project
• Mul=linguism
• 2nd scenario:
• op=miza=on of storage in the reuse of valued waste
• generic applica=ons
44
Future work
45
Thank you for your aden-on !
me : [email protected] http://www-sop.inria.fr/members/Freddy.Limpens
my advisors : Fabien Gandon : [email protected] Michel Buffa : [email protected]
ISICIL team : http://isicil.inria.fr