Keynote at KESW2014 (http://2014.kesw.ru/) about Dataset Characterisation and Linking.

1. Turning Data into Knowledge – profiling and interlinking Web datasets
Stefan Dietze
L3S Research Center
- KESW2014 -
30/09/14
1
Stefan Dietze
KESW2014

2. KESW2014
Recent work on Linked Data exploration/discovery/search
 Entity interlinking & dataset interlinking recommendation
 Dataset profiling
 Data consistency & conflicts
Research areas
 Web science, Information Retrieval, Semantic Web & Linked
Data, data & knowledge integration (mapping, classification,
interlinking)
 Application domains: education/TEL, Web archiving, …
Some projects
Introduction
http://www.l3s.de/
30/09/14 2
 See also: http://purl.org/dietze
Stefan Dietze

3. KESW2014
…why are there so few datasets actually used?
Date reuse and in-links focused on trusted „reference graphs“ such as DBpedia, Freebase etc
Long tail of LD datasets which are neither reused nor linked to (LOD Cloud alone 300+ datasets, 50 bn triples)
Explanations?
Linked Data is awesome, but...
30/09/14
„HTTP-accessibility“ (SPARQL, URI-dereferencing)
„Structure“ & „Semantics“ (=> shared/linked vocabularies)
„Interlinked“
„Persistent“
Hm, really?
Stefan Dietze
3

4. KESW2014
Linked data is more diverse (and messy) than we think
SPARQL endpoint availability over time [Buil-Aranda et al 2013]
Accessibility of datasets?
Less than 50% of all SPARQL endpoints actually responsive at given point of time [Buil-Aranda2013]
“THE” SPARQL protocol? No, but many variants & subsets
“Semantics”, links, quality?
…data accuracy (eg DBpedia)? [Paulheim2013]
…vocabulary reuse? [D’AquinWebSci13]
…schema compliance (RDFS, schemas) [HoganJWS2012]
Stefan Dietze
SPARQL Web-Querying Infrastructure: Ready for Action?, Carlos Buil-Aranda, Aidan Hogan, Jürgen Umbrich Pierre-Yves Vandenbussch, International Semantic Web Conference 2013, (ISWC2013).
Assessing the Educational Linked Data Landscape, D’Aquin, M., Adamou, A., Dietze, S., ACM Web Science 2013 (WebSci2013), Paris, France, May 2013.
Type Inference on Noisy RDF Data, Paulheim H., Bizer, C. Semantic Web – ISWC 2013, Lecture Notes in Computer Science Volume 8218, 2013, pp 510-525
An empirical survey of Linked Data conformance. Hogan, A., Umbrich, J., Harth, A., Cyganiak, R., Polleres, A., Decker., S., Journal of Web Semantics 14, 2012
30/09/14
4

5. KESW2014
What about data consistency?
Analyzing Relative Incompleteness of Movie Descriptions in the Web of Data: A Case Study, Yuan, W., Demidova, E., Dietze, S., Zhu, X., International Semantic Web Conference 2014 (ISWC2014)
30/09/14
Stefan Dietze
5

6. KESW2014
Too many/diverse datasets, too little knowledge
Stefan Dietze
30/09/14
?
?
?
?
?
?
Topics? Which datasets are useful & trustworthy for case XY (eg „learning about the solar system“) ? Which topics are covered?
Types? Which datasets describe statistics, videos, slides, publications etc?
Quality? Currentness, dynamics, accessability/reliability, data quantity & quality?
6

7. KESW2014
db:Astro. Objects
Dataset Metadata
Stefan Dietze
30/09/14
BIBO
AAISO
FOAF
contains
Entity & dataset disambiguation & linking [ESWC13]
Topic profile extraction [WWW13, ESCW14]
db:Astronomy
db:Astro. Objects
Dataset Catalog/Registry
yov:Video
po:Programme
BBC Programme
<po:Programme …> <po:Series>Wonders of the Solar System</.> <po:Actor>Brian Cox</…> </po:Programme…>
<yo:Video …> <dc:title>Pluto & the Dwarf Planets</dc:title> … </yo:Video…>
Yovisto Video
bibo:Fil
bibo:Fi
bibo:Film
Schema mappings [WebSci13]
Data mapping, linking and profiling
7

8. KESW2014
Schemas/vocabularies on the Web: XKCD 927
Stefan Dietze
30/09/14
https://xkcd.com/927/
schemas & vocabularies
8

9. KESW2014
Schema assessment and mapping
Co-occurence of data types (in 146 datasets: 144 Vocabularies, 588 highly overlapping types, 719 Properties)
Assessing the Educational Linked Data Landscape, D’Aquin, M., Adamou, A., Dietze, S., ACM Web Science 2013 (WebSci2013), Paris, France, May 2013.
po:Programme
sioc:Item
30/09/14
yov:Video
?
Stefan Dietze
9

10. KESW2014
typeX
typeX
Schema assessment and mapping
Co-occurence of data types (in 146 datasets: 144 Vocabularies, 588 highly overlapping types, 719 Properties)
Co-occurence after mapping into most frequent schemas (201 frequent types mapped into 79 classes)
Assessing the Educational Linked Data Landscape, D’Aquin, M., Adamou, A., Dietze, S., ACM Web Science 2013 (WebSci2013), Paris, France, May 2013.
bibo:Film
bibo:Document
po:Programme
sioc:Item
30/09/14
foaf:Document
yov:Video
typeX
10

11. KESW2014
Application: LinkedUp Data Catalog
in a nutshell
 RDF (VoID) dataset catalog: browse &
query distributed datasets
 Federated queries using type mappings
 Live information about endpoint
accessibility
Stefan Dietze 30/09/14
11
http://data.linkededucation.org/linkedup/catalog/
http://datahub.io/group/linked-education
DBpedia categories

12. KESW2014
Stefan Dietze
30/09/14
contains
yov:Video
po:Programme
BBC Programme
<po:Programme …> <po:Series>Wonders of the Solar System</.> <po:Actor>Brian Cox</…> </po:Programme…>
<yo:Video …> <dc:title>Pluto & the Dwarf Planets</dc:title> … </yo:Video…>
Yovisto Video
Towards profiling: dataset disambiguation/linking
?
Relatedness of entities, meaningfulness of paths? [ESWC13]
Extraction of “topics” & relatedness of datasets [ESWC14]
?
?
?
14
db:Astro. Objects
db:CartoonCharacters
?

13. KESW2014
Stefan Dietze
30/09/14
contains
yov:Video
po:Programme
BBC Programme
<po:Programme …>
<po:Series>Wonders of the Solar System</.>
<po:Actor>Brian Cox</…>
</po:Programme…>
<yo:Video …>
<dc:title>Pluto & the Dwarf Planets</dc:title>
…
</yo:Video…>
Yovisto Video
Combining a co-occurrence-based and a semantic measure for entity linking, B. P. Nunes, S. Dietze, M.A. Casanova, R. Kawase, B. Fetahu, and W. Nejdl., ESWC 2013 - 10th Extended Semantic Web Conference, (May 2013).
db:Pluto (Dwarf Planet)
db:Astrono- mical Objects
db:Sun
db:Astronomy
Computation of connectivity scores between entities
Combination of a (i) semantic (graph-based) connectivity score (SCS) with (ii) a Web co-occurence-based measure (CBM) (similar to NGD)
For (i): adaptation of Katz-Index from SNA for (linked) data graphs (considering path number and path lengths of transversal properties)
SCS = 0.32
CBM = 0.24
15
Dataset disambiguation/linking

14. KESW2014
Entity linking: evaluation
30/09/14
16
Stefan Dietze
 Evaluation based on USA Today News items (80.000 entity pairs)
 Manually created gold standard (1000 entity pairs)
 Baseline: Explicit Semantic Analysis (ESA) => CBM/SCS: „relatedness“; ESA: „similarity“
Precision/Recall/F1 for SCS, CBM, ESA.
Combining a co-occurrence-based and a semantic measure for entity linking, B. P. Nunes, S. Dietze, M.A. Casanova, R. Kawase, B. Fetahu, and W. Nejdl., ESWC 2013 - 10th Extended Semantic Web Conference, (May 2013).

15. KESW2014
„SCS Connector“ demo
http://lod2.inf.puc-rio.br/scs/SemConnectivities
SCS Connector – Quantifying and Visualising Semantic Paths between Entity Pairs, Nunes, B. P., Herrera, J. E. T., Taibi, D., Lopes, G. R., Casanova, M. A., Dietze, S., Demo Paper at 11th Extended Semantic Web Conference (ESWC2014), Heraklion, Crete, Greece, (2014. – *BEST ESWC2014 DEMO AWARD*
17
Stefan Dietze
30/09/14

16. KESW2014
Dataset Metadata
db:Astronomy
db:Astro. Objects
Dataset Catalog/Registry
yov:Video
<yo:Video …>
<dc:title>Pluto & the Dwarf Planets</dc:title>
…
</yo:Video…>
Yovisto Video
Extracting representative (DBpedia) categories („topic profile“) & entities for arbitrary datasets
Sounds easy? But how to do that for 300+ datasets with < 50 bn triples?
Scalability vs representativeness: sampling & ranking for good scalability/accuracy balance [ESWC2014] (applied to all responsive LOD datasets)
A Scalable Approach for Efficiently Generating
Structured Dataset Topic Profiles, Fetahu, B., Dietze, S., Nunes, B. P., Casanova, M. A., Nejdl, W., 11th Extended Semantic Web Conference (ESWC2014), Crete, Greece, (2014).
Dataset profiling: what‘s the data about?
18
Stefan Dietze
30/09/14
db:Pluto (Dwarf Planet)

17. KESW2014
Efficient dataset profiling: method
1.Sampling of resource instances (random sampling, weighted sampling, resource centrality sampling)
2.Entity and topic extraction (NER via DBpedia Spotlight, category mapping and expansion)
3.Normalisation and ranking (using graphical- models such as PageRank with Priors, HITS with Priors and K-Step Markov)
Result: weighted dataset-topic profile graph
A Scalable Approach for Efficiently Generating
Structured Dataset Topic Profiles, Fetahu, B., Dietze, S., Nunes, B. P., Casanova, M. A., Nejdl, W., 11th Extended Semantic Web Conference (ESWC2014), Crete, Greece, (2014).
19
Stefan Dietze
30/09/14

18. KESW2014
Dataset profiling: exploring LOD datasets/topics in a nutshell
http://data-observatory.org/lod-profiles/
Automatic extraction of dataset “topics” [ESWC2014] => RDF/VoiD dataset profiles
Visualisation & exploration of dataset-topic graph (datasets, topics, relationships)
Includes all (responsive) datasets of LOD Cloud
20
Stefan Dietze
30/09/14

19. KESW2014
Dataset profiling: evaluation
NDCG (averaged over all datasets) .
Datasets & Ground Truth
Yovisto, Oxpoints, LAK Dataset, Semantic Web Dogfood
Crowd-sourced topic indicators from datasets (keywords, tags)
Manual mapping to entities & category extraction (ranking according to frequency) Baselines
1) LDA, 2) tf/idf (applied to entire datasets)
Topic extraction according to our approach, weighting/ranking based on term weight Measure
NDCG @ rank l
Performance (time/NDCG) for different sampling strategies/sizes etc
21
Stefan Dietze
30/09/14

20. KESW2014
30/09/14
What (dataset) have these categories in common?
dbp:Category:1955_births
dbp:Category:People_from_London
dbp:Category:Buzzwords
dbp:Category:Semantic_Web
dbp:Category:Web_Services
dbp:Category:HTTP
dbp:Category:Unitarian_Universalists
dbp:Category:World_Wide_Web
dbp:Category:Royal_Medal_winners
Stefan Dietze
22
?
?

21. KESW2014
30/09/14
Diversity of category profile for a single publication
Berners-Lee, Tim; Hendler, James, Ora Lassila (2001). "The Semantic Web". Scientific American Magazine.
foaf:Person
foaf:Document
dbp:Tim_Berners-Lee
dbp:Category:1955_births
dbp:Category:People_from_London
dbp:Category:Buzzwords
dbp:Semantic_Web
dbp:Category:Semantic_Web
dbp:Category:Web_Services
dbp:Category:HTTP
dbp:Category:Unitarian_Universalists
first-level categories (dcterms:subject)
dbp:Category:World_Wide_Web
dbp:Category:Royal_Medal_winners
Stefan Dietze
DBLP
23

22. KESW2014
30/09/14
http://data-observatory.org/led-explorer/
Type specific views on datasets/ categories
“Document” (foaf:document)
“Person “ (foaf:person)
“Course” (aaiso:course)
Currently applied to datasets in LinkedUp Catalog only (as schema mappings already available here)
Type-specific exploration of dataset categories
Stefan Dietze
Exploring type-specific topic profiles of datasets: a demo for educational linked data, Taibi, D., Dietze, S., Fetahu, B., Fulantelli, G., Demo at International Semantic Web Conference 2014 (ISWC2014)
24

23. KESW2014
data.l3s.de – the L3S DataHub

24. KESW2014
KEYSTONE & PROFILES 2014
30/09/14
27
Stefan Dietze
http://www.keystone-cost.eu/
KEYSTONE: semantic keyword-based search on structured data sources (2013-2017)
Research network focused on distributed search, dataset profiling, to Semantic Web, Databases, etc.
Open to new members (beyond Europe)
http://www.keystone-cost.eu/profiles
http://www.ijswis.org/?q=node/51/
PROFILES2014 - Dataset PROFIling & fEderated Search for Linked Data
Workshop collocated with ESWC2014
IJSWIS Special Issue on … LD search & profiling
Deadline 8 December 2014

25. KESW2014
Summing up
Summary
Increasing amounts of data => require knowledge about nature and relationships of datasets
Profiling: scalable methods for extracting dataset metadata
Interlinking: connectivity of entities or datasets What about LD evolution?
In RDF graphs (eg LOD Cloud), „all“ nodes are connected
Impact of evolution on preservation, linking and enrichment?
Which parts of datasets to preserve (entity „neighbourhood“)? => semantic relatedness /relevance/entity retrieval
Link correctness in evolving LD?
….
30/09/14
29
Stefan Dietze

26. KESW2014
Спасибо! Thank You!
WWW See also (general)
 http://purl.org/dietze
 http://linkedup-project.eu
 http://duraark.eu
 http://data.l3s.de See also (data)
 http://data.l3s.de
 http://data.linkededucation.org
http://lak.linkededucation.org
30/09/14
30
Stefan Dietze
Besnik Fetahu (L3S)
Elena Demidova (L3S)
Bernardo Pereira Nunes (PUC Rio)
Marco Casanova (PUC Rio)
Luiz Andre Paes Leme (PUC Rio)
Giseli Lopes (PUC Rio)
Davide Taibi (CNR, IT)
Mathieu d’Aquin (Open University, UK)
and many more…
Acknowledgements