Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
Session 0.0 poster minutes madness
1. SEMANTiCS 2017, Amsterdam, The Netherlands, September 11-14, 2017
Towards a Semantic Outlier Detection
Framework in Wireless Sensor Networks
Iker Esnaola-Gonzalez
2. Is there a country that publishes formal key registers as linked data?
9. Semantic Health Notice
Come and get the cure ! (free trial)
http://www.irisa.fr/LIS/ferre/sparklis/
Sébastien Ferré
● Ground: SPARQL endpoints
● Patients: RDF-hungry people
● Symptoms: paralysis (can’t write!), syntax errors,
schema mismatch, empty results
➔ FRUSTRATION and HUNGER !
● Diagnostic: SPARQL queries are toxic to humans
●
Treatment: use SparklisSparklis to avoid direct contact
composition: natural language, full guidance, expressive
power, compliance to standards (no empty-results)
P&D 204
SPARKLIS
10.
11. Semantic mining of respiratory health knowledge from
scientific literature and clinical guidelines
Sergio Consoli, Xiao Ming Zhou, Wei-Shun Bao, Declan P. Kelly, Vincent Lou
Philips Research: [name.surname]@philips.com
• Application of NLP, data mining, and advanced text analytics into advanced AI systems for semantically mining
information from scientific literature and clinical guidelines to extract new knowledge on the respiratory health domain
• Curated repository of selected medical documents into knowledge-base in respiratory health
12. • Goal feeding and controlling connected respiratory health devices according the gathered knowledge and providing
intuitive visualization interfaces for user understanding
Scope 1: data bootstrap via scientific literatures to extract knowledge
Scope 2: provide relevant
feedback and insights from data
Challenges
• Repository of biomedical scientific literature
– Large volume & Accessibility
• Data Processing and Interpretation
– Performance VS Speed
• Knowledgebase
– Data-linking of concepts and relationships to ensure both syntactic and semantic interoperability
• Information Querying and Retrieval
– ML & prediction models for deriving rules and facts, and providing feedbacks/insights/recommendations
13. Is Semantic Markup really
helping websites improve
traffic?
Build Your Own Knowledge Graph and make
your content easy to be found.
WordLift team wordlift.io
14. Key Findings from our research
1. After 3 Months organic visits went up by 12.13%
2. Google is faster in bringing new users (+18.47%
increased of sessions from Google only)
3. Enriched article compared to other pages performed
2.4 times better in terms of pageviews and sessions
4. Average time spent went up 17.3% for enriched
articles
5. Average session duration improved by 13.75%
Yes it does!
15. “Findability” is not simply SEO
BEFORE and AFTER
Get Your Free Demo! wordlift.io/welcome
16. Towards Open Data Mashups for Data Journalism
Fajar J. Ekaputra, Niina Maarit Novak, Elmar Kiesling, Peb R. Aryan, Ba-Lam Do, Tuan-Dat Trinh, A Min Tjoa
Linked Data Lab, Institute of Software Technology and Information Systems, TU Wien, Austria
Prototype:Open Data Mashups for Data Journalism (ODMOJO):
17. PRESS: A Publication REpository Semantic System
• PRESS is an open-source publication system that exploits semantic
technologies, in order to cover the needs of both individuals and
organizations
• Fast data entry, user-based repository
• Advanced query capabilities
• Integration with other systems
• Commercial: designed mainly for experts, not extensible
• Existing open-source: difficulty in expressing complex queries, rely on rigid relational schemas
FORTH, Institute of Computer Science
VS. Other Approaches
Ioannis Chrysakis, Emmanouil Dermitzakis, Giorgos Flouris, Theodore Patkos and Dimitris Plexousakis
Learn more at stand #2
18. Vision and market gap
• Motivation
– Comments are becoming an integral part
of the Web, and not only the Social Web
– Yet, comment management is largely ad-
hoc and inefficient
• Grand vision: interconnect, search,
store, retrieve, evaluate comments that
exist on the Web
Methinks
Methinks Application
• Advantages for consumers/end-users
– Become well-informed in just a glance
– Quickly identify non-standard information that can be
found only in comments
– Browse based on arguments, not on words
• Advantages for analysts (admins)
– A cheap way to integrate fancy, comment-related
functionality in one’s web site
– A measurable way to extract conclusions (analytics) in
online discussions
Contact: Giorgos Flouris (fgeo@ics.forth.gr)Learn more at stand #3
19. Information is archived in a way
disregarding its meaning and context
Retrieval systems search, analysis
Retrieval systems visualization
Information is archived in a way disregarding
itsmeaning and context
Retrievalsystems search, analysis
Retrieval systems visualization options
Problem & VIScover SolutionUncovering the Hidden in Large Knowledge Graphs
1
How to connect the dots in my Knowledge Graph?
• How to get meaningful insight?
• Which groups of objects connect
to each other?
• How to build sophisticated
queries without syntax?
• How about data integrity or
anomalies?
naive graph
rendering fails
SPARQL, cypher, etc.
require data insight
20. Information is archived in a way
disregarding its meaning and context
Retrieval systems search, analysis
Retrieval systems visualization
Information is archived in a way disregarding
itsmeaning and context
Retrievalsystems search, analysis
Retrieval systems visualization options
Problem & VIScover SolutionSemSpect: Visualizing and Querying Knowledge Graphs
2
Poster &
Wed.
10:45
http://panama.semspect.de
21. Michael Färber and Achim Rettinger, KIT: A Statistical Comparison of Current Knowledge Bases
22. Now: In a Crunch, so CrunchBase
Michael Färber, Carsten Menne, Andreas Harth: “A Linked Data Wrapper for CrunchBase”
23. 1WIS
Challenge: Form an Educated Opinion for Online Products
Twilight (2008): 1545 reviews on IMDB.com, 3,053 reviews on Amazon.com.
Which reviews to read?
Example: Is the movie “Twilight” any good?
Horrible Story, Horrible ActingGreat Story, Great Acting
Sorted by Usefulness
Grouped by Aspects
VS.
Proposed solution: Semantically group reviews to help users!
25. 1
● Statement-level metadata allows to store
fine-grained
○ traceability and provenance information
○ license and access rights,
○ data trustworthiness and confidence
scores
for every single fact in the knowledge graph
● Different MRMs have been proposed, varying
in structure, performance
benefits/disadvantages
● Hard to choose appropriate format
beforehand
MaSQue: An Approach for Flexible Metadata
Storage and Querying in RDF
Knowledge Integration and Linked Data Technologies (AKSW/KILT)
at InfAI & Leipzig University
Johannes Frey, Sebastian Hellmann
Metadata Representation Models
27. Network-based Knowledge Graph Assessment
What is it about?
Real-world entities and their relations
Information retrieval
Effectiveness depends on quality
1Rörden, Revenko, Haslhofer, Blumauer:
Network-based Knowledge Graph Assessment.
28. Network-based Knowledge Graph Assessment
What is our approach?
Network metrics
Domain-specific corpus
Combine and compare!
2Rörden, Revenko, Haslhofer, Blumauer:
Network-based Knowledge Graph Assessment.
29. ‹#› Het begint met een idee
A STUDY OF INTENSIONAL CONCEPT DRIFT
IN TRENDING DBPEDIA CONCEPTS
Albert Meroño-Peñuela, Efstratios Kontopoulos,
Sándor Darányi, Ioannis Kompatsiaris
Causes?
• Interaction between ordinary humans and systems
• Ontology evolution patterns of use
34. Introduction
DALICC stands for Data Licenses Clearance Center. The
project‘s aim is to develop a software framework that
significantly reduces the costs of license clearance in the
creation of derivative (data) works.
DALICC follows a deontic approach to express machine-
readable permissions, prohibitions and duties defined in a
license. DALICC utilizes the ODRL 2.0 ontology in
combination with CCRel. Additionally DALICC utilizes an
extended set of properties that are necessary to capture the
full semantic spectrum of legal expressions defined in a
license text.
By doing so DALICC allows to represent property rights
policies in a legally valid and machine-processable way.
Partners:
• UAS St. Pölten (Lead) – Tassilo Pellegrini, Andrea Schönhofer, Peter Judmaier, Stefanie Größbacher
• Vienna University of Economics & Business – Simon Steyskal, Sabrina Kirrane, Axel Polleres
• University of Innsbruck – Oleksandra Panasiuk, Anna Fensel
• Höhne, In der Maur & Partner Rechtsanwälte OG – Markus Dörfler
• Semantic Web Company GmbH – Thomas Thurner, Victor Mireles-Chavez, Kurt Moser
Contact: www.dalicc.net
Acknowledgment: DALICC is funded by the Austrian Federal Ministry of Transport, Innovation and Technology
(BMVIT) under the program "ICT of the Future" between November 2016 - October 2018. More information
https://iktderzukunft.at/en/
DALICC Vocabulary – REL Extension
The DALICC vocabulary extensions evolved from an in depth
analysis of 15 standard licenses (including CC, APACHE, BSD,
GPL MIT) under the supervision of legal experts. The following
expressions need to be added to cover the full semantic
spectrum of copyright statements:
PROPERTIES AFFECTING THE ASSET PROPERTIES AFFECTING THE LICENCE
dalicc:charge dalicc:addStatement
dalicc:sublicense dalicc:attributionNotice
dalicc:promote dalicc:attachOffer
dalicc:publish dalicc:chargeOffer
dalicc:irrevocable
dalicc:modificationNotice
dalicc:noWarrantyNotice
dalicc:patentFree
dalicc:patentNotice
dalicc:perpetual
dalicc:royaltyFree
dalicc:worldwide
The examples illustrate the RDF/Turle syntax of BSD 3.0, CC-BY
and APACHE using the DALICC properties (bold).
Example 3: APACHE
@prefix odrl:<http://www.w3.org/ns/odrl/2/> .
@prefix : <https://dalicc.net/license-finder> .
@prefix dalicc: <https://dalicc.poolparty.biz/DALICCVocabulary>.
@prefix rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix cc:<http://creativecommons.org/ns#> .
@prefix dct: <http://purl.org/dc/terms/> .
@prefix foaf: <http://xmlns.com/foaf/0.1/> .
@prefix rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
:APACHE_2.0 a odrl:Policy;
odrl:permission [
a odrl:Permission;
odrl:target
<http://purl.org/dc/dcmitype/Software>,
<http://purl.org/dc/dcmitype/Sound>,<http://purl.
org/dc/dcmitype/Text>,<http://purl.org/dc/dcmity
pe/Image>,<http://purl.org/dc/dcmitype/MovingI
mage>,<http://purl.org/dc/dcmitype/Dataset>;
odrl:action odrl:present, odrl:display, odrl:derive;
odrl:duty [
a odrl:Duty;
odrl:action cc:Notice;
]
];
odrl:permission [
a odrl:Permission;
odrl:target
<http://purl.org/dc/dcmitype/Software>,<http://p
url.org/dc/dcmitype/Sound>,<http://purl.org
/dc/dcmitype/Text>,<http://purl.org/dc/dcmitype
/Image>,<http://purl.org/dc/dcmitype/Movin
gImage>,<http://purl.org/dc/dcmitype/Dataset>;
odrl:action odrl:reproduce, odrl:distribute,
dalicc:sublicense, dalicc:addStatement,
dalicc:chargeOffer, dalicc:patentFree;
odrl:duty [
a odrl:Duty;
odrl:action dalicc:modificationNotice, cc:Notice,
dalicc:patentNotice, dalicc:atributionNotice,
dalicc:noWarrantyNotice, cc:ShareAlike,
dalicc:perpetual, dalicc:royaltyFree,
dalicc:irrevocable, dalicc:worldwide
]
];
odrl:prohibition [
a odrl:Prohibition;
odrl:target
<http://purl.org/dc/dcmitype/Software>,<http://p
url.org/dc/dcmitype/Sound>,<http://purl.org/dc/d
cmitype/Text>,<http://purl.org/dc/dcmitype/Image
>,<http://purl.org/dc/dcmitype/MovingImage>,
<http://purl.org/dc/dcmitype/Dataset>;
odrl:action odrl:ensureExclusivity, dalicc:charge;
].
Example 1: BSD 3.0
@prefix odrl:<http://www.w3.org/ns/odrl/2/> .
@prefix : <https://dalicc.net/license-finder> .
@prefix dalicc: <https://dalicc.poolparty.biz/DALICCVocabulary>.
@prefix rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix cc:<http://creativecommons.org/ns#> .
@prefix dct: <http://purl.org/dc/terms/> .
:licBSD-3-Clause a odrl:Policy;
odrl:permission [
a odrl:Permission;
odrl:target <http://purl.org/dc/dcmitype/Software>;
odrl:action odrl:mofify, odrl:distribute, odrl:reproduce;
odrl:duty [
a odrl:Duty;
odrl:action cc:Notice, dalicc:nowarrantyNotice;
]
];
odrl:prohibition [
a odrl:Prohibition;
odrl:target <http://purl.org/dc/dcmitype/Software>;
odrl:action dalicc:promote
];
dct:title "The 3-Clause BSD License"@en ;
dct:alternative "BSD-3-Clause";
dct:source <https://opensource.org/licenses/BSD-3-Clause>.
Example 2: CC-BY
@prefix odrl:<http://www.w3.org/ns/odrl/2/> .
@prefix : <https://dalicc.net/license-finder> .
@prefix dalicc: <https://dalicc.poolparty.biz/DALICCVocabulary>.
@prefix rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix cc:<http://creativecommons.org/ns#> .
@prefix dct: <http://purl.org/dc/terms/> .
@prefix foaf: <http://xmlns.com/foaf/0.1/> .
:CC-BY_4.0 a odrl:Policy;
odrl:permission [
a odrl:Permission;
odrl:target
<http://purl.org/dc/dcmitype/Dataset>,<http://purl.org/d
c/dcmitype/Sound>,<http://purl.org/dc/dcmitype/Text>,<
http://purl.org/dc/dcmitype/Image>,<http://purl.org/dc/
dcmitype/MovingImage>;
odrl:action odrl:distribute, odrl:reproduce,odrl:extract,
odrl:derive, odrl:present;
odrl:duty [
a odrl:Duty;
odrl:action cc:SourceCode, dalicc:royaltyFree,
dalicc:irrevocable, dalicc:worldwide, cc:Notice,
dalicc:noWarrantyNotice, dalicc:modificationNotice,
cc:attributionName
]
];
odrl:prohibition [
a odrl:Prohibition;
odrl:target
<http://purl.org/dc/dcmitype/Dataset>,<http://purl.org/dc
/dcmitype/Sound>,<http://purl.org/dc/dcmitype/Text>,
<hattp://purl.org/dc/dcmitype/Image>,<http://purl.org/dc
/dcmitype/MovingImage>;
odrl:action odrl:ensureExclusivity, dalicc:sublicense
];
dct:title "Attribution 4.0 International"@en ;
dct:alternative "CC BY 4.0";
dct:publisher "Creative Commons";
foaf:logo <http://i.creativecommons.org/l/by/4.0/88x31.png> ;
dct:source <http://creativecommons.org/licenses/by/4.0/> ;
cc:legalcode
"""https://creativecommons.org/licenses/by/4.0/legalcode"""@en.
Future Work
The DALICC Framework will consist of four components:
• License Composer: Lets you create customized licenses
• License Library: Lets you choose from a set of standard licenses
• License Annotator: Provides you with a machine-readable and human-readable version of
your license
• License Negotiator: Checks compatibility, detects conflicts and supports conflict resolution
Data Set
N License N
License
{A … N} Derivative Work
Data Set
… License …
Data Set
B License B
Data Set
A License A
License
Composer
License
Annotator
License
Library
DALICC Framework
License
Negotiator
consults
tags audits proposes
• Compatibility
• Conflict Detection
• Conflict Resolution
35. ‹#› Het begint met een idee
CEDAR: THE DUTCH HISTORICAL CENSUSES
(1795-1971) AS LINKED OPEN DATA
Albert Meroño-Peñuela, Ashkan Ashkpour,
Christophe Guéret, Stefan Schlobach
1st historical census data as Linked Data
RDF Data Cube with
• 6.8M observations
• Internal links: sex, marital status, occupation
position, housing type, residence status
• External links
• Geographical: 2.7M
• Occupations: 350K
• Belief: 250K
• Linked to HISCO, ICONCLASS, Dutch Ships and
Sailors, DBpedia, gemeentegeschiedenis.nl,
GeoNames
• New variables and classifications
• Housing type
• Lower level of municipal areas (kom, wijk)
• Residence status
36. ‹#› Het begint met een idee
http://www.licr.io
http://purl.org/licr/vocab
Bottom-up codes
from:
• NAPP
• IPUMS
• HL7
37. ‹#› Het begint met een idee
HISCO – Historical International Standard
Classification of Occupations
9 major groups
1675 occupation
classes
Labels in 14
languages
38. SEMANTiCS 2017, Amsterdam, The Netherlands, September 11-14, 2017
The EEPSA Ontology
The Energy Efficiency Prediction Semantic Assistant Ontology
Iker Esnaola-Gonzalez
https://w3id.org/eepsa
40. From diverse and
complex solutions...
The Linked Web APIs Ontology
- semantic Web API description model -
...to a unified
solution.
FOAF
Hydra
SADI
WSMO
SAWSDL
OpenAPI
apis.json
The Linked Web APIs
Ontology
WADL
WSDL
“A lightweight model for
description of information
related to Web APIs,
mashups, developers and
API providers.”
WADL
MSM
OWL-S
HTML
See: http://linked-web-apis.fit.cvut.cz/ns/core/
41. (Non-)Functional props:
“what functionalities a Web
API or mashup offers.”
Provenance info:
“Who created What
and How”
Temporal information:
“the time a Web API or a
mashup was created.”
Technical props:
“supported protocols,
formats, endpoints, etc.”
Over 11K Web APIs,
7K mashups and
7K developer profiles
First and largest Linked Data
dataset with semantic Web API
descriptions!
The Linked Web APIs Dataset
- Web APIs meet Linked Data -
More info at: http://linked-web-apis.fit.cvut.cz/
42. PRODUCT PASSPORT OF YOUR GARMENT
Produced By: Mud Jeans
Product ID: RM1600016
Production Step Produced By Production Date Composition
Textiles Collected Mud Jeans 01-05-2016 Discarded Jeans 100%
Recycled Fibers Recover 10-05-2016 Denim Recycled 100%
Garment Produced Yousstex 15-09-2016
Denim Recycled 20%;
Cotton 75%
Activity Agent
Product
Linked Data
Circular Economy
Textile Use Case
Elke Sauter & Martijn Witjes
46. Talk of Europe
Laura Hollink, Astrid van Aggelen, Jacco van Ossenbruggen
The debates of the European
Parliament as Linked Open Data*
Centrum Wiskunde & Informatica (CWI)
Amsterdam, The Netherlands
Talk of Europe
made by
the ECLinks:
Motivation:
Access is crucial for EU residents
to make informed votes and hold
Members of Parliament accountable.
Access is necessary
for researchers in
party position, ideology,
issue selection, and
emphasis.
What?
1. scrape http://www.europarl.europa.eu/
2. convert to RDF
3. enrich with links to
external sources
Why?
Access is crucial for EU residents
to make informed votes and hold
Members of Parliament accountable.
Access is necessary
for researchers in
party position, ideology,
issue selection, and
emphasis.
47. Example query: The number
of speeches of each country
per year.
session
session day
dcterms:partOf
item on the
agenda
partOf
speech
partOf
...This is indeed a
homecoming...@en
spoken
text
...Dette er i høj grad
en hjemkomst...@da
translated
text
speaker
speaker
member of
parliamentrdf:type
political function
president role
2002
beginningEuropean
Parliament
institution
2004
end
political function
substituterole
1989
beginning
Committee on
budgetary control
institution
1992
end
sub-properties
of text
design pattern:
n-ary relation
spokenAs
redundant but convenient
additional property
foaf:name
....
re-use of Dublin Core
and FOAF vocabularies
Vocabulary:
Published as:
Astrid van Aggelen, Laura Hollink, Max Kemman, Martijn
Kleppe, Henri Beunders. The debates of the European
Parliament as Linked Open Data. Semantic Web 8(2), IOS
Press, 2017
Website: http://talkofeurope.eu/
SPARQL endpoint: http://purl.org/linkedpolitics/sparql/
Acknowledgements: