Date post: | 28-Jul-2015 |
Category: |
Science |
Upload: | erik-mannens |
View: | 247 times |
Download: | 3 times |
ESWC 2015: EU Project Networking
Session
3rd June 2015 (14:00 – 17:00)
Room Adria I/II, Grand Hotel Bernardin, Portoroz
EU Project Networking Session 2015: Organizers
Frédérique Segond (Viseo, Grenoble, France)
Sergio Consoli (STLab ISTC-‐CNR, Italy)
Jun Zhao (Lancaster University, UK) Erik Mannens (iMinds-‐Ghent Univ.-‐MMLab, Be)
EU Project Networking Session 2015: Purpose
To provide EU projects with the ability q to connect with each other and engage in discussions about their
respecNve research and development, q to establish opportuni2es for knowledge and technology sharing, q to iden2fy complementary ac2vi2es and goals which can form the
basis for § future collaboraNons, § research proposals, § researcher exchange, and/or § joint parNcipaNon at events/iniNaNves
EU Project Networking Session 2015: Program
14:05-‐14:15 à Session Opening 14:15-‐14:50 à Madness-‐Presenta2ons 15:00-‐16:00 à Projects’ Posters
16:00-‐17:00 à Thema2c Tables 16:55-‐17:00 à Session Wrap-‐Up
EU Project Networking Session 2015: One Minute Madness
14:15-‐14:50 à Madness-‐Presenta2ons q You have two minutes to present your project! Strict 6me window!
q Flash highlight of the project
q Just answer the Five Ws: WHO; WHAT; WHEN; WHERE; WHY
q Please follow the order list we have given you, and be ready when it is your turn!
1
LEO
Linked Open Earth Observation Data for Precision Farming (LEO, http://www.linkedeodata.eu/)
Manolis Koubarakis
National and Kapodistrian
University of Athens
The project LEO
04/06/15 Manolis Koubarakis 8
• LEO studies the life cycle of linked EO data and develops tools that support it. The results of the project are the following:
– GeoTriples: a tool that extracts linked geospatial data from their native formats (e.g., shapefiles) and transforms them into RDF.
– Extension of Silk: techniques for interlinking linked geospatial and temporal data.
– LEO DSE, LEODroid, Sextant: tools for searching, visualizing and exploring linked EO data and linked geospatial data. Developed for mobile platforms (Android).
– LEOpatra: a precision farming application that exploits linked geospatial data.
2
Open Data Monitor
OpenDataMonitor
FP7-ICT-2013.4.3 SME initiative on analytics Project number: 611988
OpenDataMonitor Monitoring, analysis and visualisation of open data catalogues, hubs and repositories
What we aim to do • Gain an overview of the open data ecosystem • Analyse and visualise data catalogues using
innovative technology
How we do it • Harvest and harmonize multilingual metadata
from data catalogues • Make information available at municipal,
national, and pan-European levels
project.opendatamonitor.eu
Open Data Monitor: opendatamonitor.eu
3
MARIO
MARIO: Managing active and healthy Aging with use of caRing servIce rObots Diego Reforgiato Recupero, Aldo Gangemi, Misael Mongiovi, Stefano Nolfi, Andrea G. Nuzzolese, Valentina
Presutti, Massimiliano Raciti, Thomas Messervey, Dympna Casey, Vincent Dupourque, Geoff Pegman, Alexandros Gkiokas, Andy Bleaden, Antonio Greco, Christos Kouroupetroglou, Siegfried Handschuh
www.mario-project.eu
Started in February 2015
Robot semantics based on Semantic Web practices and technologies: Linked Data principles, RDF, SPARQL, RIF.
Mario Ontology Network (MON) will reuse and extend the Ontologies for Robotics and Automation. MON will evolve over time by integrating ontologies emerging from interaction with assisted humans, sensors or with other robots.
Semantic Web-based machine reading/listening in robots. FRED, will be extended and improved for dealing with c o n t e x t - b a s e d g r o u n d i n g a n d interpretation of natural language input.
Ability to advance robot knowledge by learning new ontology patterns from its experience with users and the robot network in place. New emerging patterns and expressions are fed back to the robot’s cognitive system in order to address emotional needs of end users in compliance with the social and behavioral objectives of MARIO.
Robot social skills: a sentiment analysis framework based on deep parsing of natural language and supported by MON will deal with moods and expression recognition providing robots.
“ E n t i t y - c e n t r i c ” k n o w l e d g e management: each entity and its relations have a public identity that provides a first “grounding” to the knowledge used by robots. Such identity is given by resolvable URIs that use simple Web and Internet protocols to p rov i de u sefu l k now ledge a s a representative of real world entities.
KOMPAI PLATFORM from Robosoft
4
PHEME
5
EGI-‐ENGAGE
EGI ENGAGE DARIAH competence centre
what ? establish science-‐oriented competence centre providing support for researchers within the humaniNes, arts and social sciences
EGI ENGAGE DARIAH competence centre
what ? establish science-‐oriented competence centre providing support for researchers within the humaniNes, arts and social sciences à further use of infrastructures in the arts & humaniNes à support development of & research in digital arts & humaniNes
EGI ENGAGE DARIAH competence centre
how ? DARIAH arts & humaniNes use case (e-‐infrastructure, educaNon+training, use cases, data)
EGI ENGAGE DARIAH competence centre
how ? DARIAH arts & humaniNes use case (e-‐infrastructure, educaNon+training, use cases, data)
language resources, AT varieNes
EGI ENGAGE DARIAH competence centre
how ? DARIAH arts & humaniNes use case (e-‐infrastructure, educaNon+training, use cases, data)
language resources, AT varieNes
à join in to increase infrastructures for the arts & humaniNes!
eveline.wandl-‐[email protected] [disseminaNon manager @ DARIAH CC]
6
x-‐LIME
LiMexLiMe – crossLingual crossMedia knowledge extraction
supported by
mainstream/ professionally produced
social/ user generated
social video, social photos
audio from TV
audio from social media
tweets, blogs, commentsreviews
TV, videos photos, images
news, annotationof audio/video
visual
auditiv
textual
Real-time content-based augmentation of Live-TV
Real-time Semantic Search across modalities
LiMexLiMe – crossLingual crossMedia knowledge extraction
Visit us at www.xlime.eu
XLiMe project
supported by
7
Facts4Workers
03.06.2015 FACTS4WORKERS – 27
Start: 1.1.2015 Duration: 4 years Partners:
• We create Smart factories for the worker.
• The worker is the smartest part of a smart factory, use his creativity and experience
• Use semantics to create workflows. We need systems which can adapt to the worker, not the other way around.
• Our semantic workflow engine combines and calls different resources to fulfill a user’s goal. It is highly adaptive and can react to knowledge entered by the user.
• Humans and machines can work together!
03.06.2015 FACTS4WORKERS – 28
8
COMSODE
The COMSODE project has received funding from the Seventh Framework Programme of the European Union in the grant agreement number 611358.
COMSODE @ ESWC2015 [email protected]
The COMSODE project • Develop a quality aware
open data pubblication platfrom (open data node)
• Provide a quality aware methodology for selecting and publishig dataset
9
ProaSense
ProaSense: The ProacNve Sensing Enterprise
• WHAT: – Support transiNon from Sensing to ProacOve Sensing Enterprises – Go from search, sensing, anOcipaOng, to proacOng. – Knowing "what might happen" and doing "what should be the best
acOon"
• HOW: – Observe-‐Orient-‐Decide-‐Act loop of situaNonal awareness – Parallel and distributed processing of high-‐velocity data from IoT – SemanNc-‐descripNon-‐supported development of proacNve real-‐Nme
applicaNons
• WHEN: 3-‐years project started in November 2013 • WHO:
ProaSense: Use Case 1: HELLA
• GOAL: – Building millions of lamps annually – Reducing scrap-‐rate and down-‐Nme
• CHALLENGES: – Heterogeneous event schemas – …messaging protocols – …event processors
• TALK TO ME ABOUT: – How to make use of semanNcs for designing stream processing pipelines? – How to scale up stream processing? – How to integrate data from IoT devices for decision support?
10
MixedEmoNons
ObjecOves -‐ large-‐scale emoNon analysis and fusion -‐ heterogeneous data: mulNlingual text, speech, image, video, social media -‐ semanNc-‐level informaNon aggregaNon and integraNon -‐ robust extracNon of social semanNc knowledge graphs for emoNon analysis
Pilots Social TV, enriching TV shows with social media emoNons, for editors and TV audience Brand ReputaOon Mgmt, tracking emoNons around brands menNons in social media, news, TV etc. Call Centres, analysing consumer and call centre operator emoNons
MixedEmoNons Social SemanNc EmoNon Analysis for InnovaNve
MulNlingual Big Data AnalyNcs Markets Gabriela Vulcu, Insight Centre for Data AnalyNcs, NaNonal University of Ireland, Galway
MixedEmoOons PlaYorm
MixedEmoNons Social SemanNc EmoNon Analysis for InnovaNve
MulNlingual Big Data AnalyNcs Markets Gabriela Vulcu, Insight Centre for Data AnalyNcs, NaNonal University of Ireland, Galway
11
Dem@Care
Demen%a Ambient Care: Mul%-‐Sensing Monitoring for Intelligent Remote Management and Decision Support
Georgios Meditskos, Ioannis Kompatsiaris
Dem@Care Facts and Figures
• CollaboraNve Project funded under FP7 ICT Call 8 • ObjecNve ICT-‐2011.5.1 “Personal Health Systems for Remote Management of Diseases, Treatment and RehabilitaNon”
• DuraNon: November 2011 – November 2015 • Budget: ~11 millions • ConsorNum of 11 partners
The Dem@Care Vision • Plaqorm for the remote care of people with demenNa
• Provide personalized care opNons • Support individuals in their daily life • Help clinicians and informal caregivers provide beser feedback
• This is achieved through mulN-‐sensor monitoring and analysis
• Key Features • ConNnuous sensor-‐based monitoring and analysis of various modaliNes
• SemanNc integraNon, analysis and interpretaNon of sensor measurements
• Personalized high-‐level descripNons of the person’s condiNon and its evoluNon
• Easy-‐to-‐use interfaces for the people with demenNa and their caregivers/clinicians
ContextDescriptor
dependency[allValuesFrom]
dul:Situation
dul:isSettingFor
describes[exactly 1]
dul:isSettingFor
leo:Event
em:Event
em:Observationem:Activity
em:Posture
em:Object
em:Action
em:Location
time:TemporalEntity
dul:Agentleo:involvedAgent
Seman%c Web Technologies in Dem@Care
12
FREME
THE FREME PROJECT
• Two year H2020 InnovaNon acNon; start February 2015
• Industry partners leading four business cases around digital content and (linked) data
• Technology development bridging language and data
• Outreach and business modelling demonstraNng moneNzaNon of the mulNlingual data value chain
13
BYTE
Big data roadmap and cross-disciplinary community for addressing societal externalities
@BYTE_EU
www.byte-project.eu
BYTE aims to assist European science and industry to gain a greater share of the big data market by 2020. In order to do so, BYTE will idenNfy measures that will help big data users to capture and amplify the posiOve externaliOes associated with big data (e.g., efficiency, innovaNon, data sharing, etc.) in a manner that enables them to diminish the associated negaOve externaliOes (e.g., privacy, data protecNon, discriminaNon, etc.).
q CoordinaNon and Support AcNon
q Mar 2014 – Feb 2017 (36 months)
q Funded by DG-‐CNCT: €2.25 million
q Grant agreement no: 619551
q 11 partners from 10 countries
Key Outcomes q Report on societal
externaliNes associated with big data
q Vision for big data in Europe
q Policy roadmap q Research roadmap q Build the big data
community Achievements of the 1st Year of work q Understand the Big data ecosystem
q DefiniNons q 10 Big Data IniNaNves
q Collect posiNve and negaNve externaliNes of big data q Case studies analysis
q 1) environmental data, 2) crisis informaNcs, 3) transport data, 4) smart ciNes data, 5) cultural data, 6) energy data and 7) health data.
q Semi-‐structured interviews and mulNdisciplinary group discussions
14
ENeL COST
EU Project Networking Session June, 3rd 2015, Portoroz
COST IS 1305 European Network of
electronic Lexicography (ENeL)
G Ontology LexicaW3C community group : The core model of Ontolex;
Figure created by John P. McCrae
G EU: European dictionary portal V 1.0 of one of the main results of the
COST Action IS 1305 ENeL
F Babelnet :Thinking lexicography outside the box :partner of COST ENeL
Partnership between ENeL and Linked Data Initiative
http://www.lider-project.eu/
https://www.w3.org/community/ontolex/ http://babelnet.org/
15
WDAqua ITN
16
LIDER
ESWC’15 EU Project Networking session 03/06/2015 Portoroz (Slovenia) 56 Jorge Gracia (UPM)
Language resources (lexicon, corpora,...)
...
LLOD generaNon
MulNlingual content metadata
LLOD (Language resources as Linked Open Data)
LOD-‐aware NLP services
MulNmedia MulNlingual Content
Content AnalyNcs
Content AnalyNcs consumers
Metadata generaNon
Content providers
Target scenario
56
ESWC’15 EU Project Networking session 03/06/2015 Portoroz (Slovenia) 57 Jorge Gracia (UPM)
Reference Architecture
Certification
Benchmarking & Validation
Discovery
LLD Linking
LLD Publishing
Metadata
Service Composition
LLD-aware Services
Licensing Provenance
Vocabularies Hosting Scalability Streaming Interoperability
Gui
delin
es a
nd S
tand
ardi
zatio
n
Multilingual Data
EU Project Networking Session 2015: Thematic Tables
16:00-‐17:00 à Thema2c Tables q The organisa6on of the tables has been based on the topics of each project
q Please volunteers to take notes J
q Table 1 : Health, EducaNon, Arts and HumaniNes § 3, 5, 11, 15
q Table 2 : (Linked)(Open)(Big)(Geo) Data § 1, 2, 4, 13
q Table 3 : MulNlingual § 6, 10, 14, 16
q Table 4 : Industry and Business § 7, 9, 12, 8
EU Project Networking Session 2015: Session Wrap-‐Up
16:55-‐17:00 à Session Wrap-‐Up q Generic Findings
§ Beser to have Round Table right a{er Minute Madness & end with (well-‐asended) poster session, which can then go on as long as needed (we had to end it now)
§ There was also a lot of “networking” beyond the EU-‐projects themselves and related demo’s were given on several tables
q Specific Results … see herea{er
Thank you very much
EU Project Networking Session 2015 –
Discussion Minutes
Table 1: Health, HumaniNes, Arts
DARIAH and EGI-‐ENGAGE • We presented the new collaboraNon between DARIAH (hsps://www.dariah.eu/) and the new EGI-‐Engage (hsps://www.egi.eu/about/egi-‐engage/).
• Discussion at the poster was lively, and interest in sharing data in the social science was shown (Andrea Maurino, Project COMSODE) and also from people not asending the session, e.g., Frank Michel (Sophia-‐AnNpolis), suggesNng further cooperaNon.
• A special topic discussed concerned the quality of the data to be processed (Christophe Lange, Bonn)
• Interest in the health data generated by Dem@Care and MARIO.
Dem@Care -‐ MARIO
• DemenNa-‐related domains • Dem@Care (FP7, from November 2011 to November 2015)
• MARIO (H2020, from February 2015 to February 2018) • Outcomes from Dem@Care can be used in MARIO as starNng points – QuesNonnaires, data, privacy/ethical guidelines, acceptability of sensors/technologies
• MARIO’s pilot data can be collected and sent to Dem@Care to be processed and analyzed by the exisNng plaqorm/components
EU Project Networking Session 2015 –
Discussion Minutes
Table 2 & 3 : (Linked) (Open) (Big) (Geo) (MulNlingual) Data
● MixedEmotions already working with LIDER; interested in PHEME, XLIME, FREME
● LIDER interested in XLIME, MixedEmotions, ENeL COST, FREME, WDAque ITN
● WDAqua ITN interested in PHEME, LIDER, ENeL COST, COMSODE
● FREME interested in MixedEmtions, ENeL COST, Open Data Monitor Project
● ProaSense discussed with PHEME (challenges in real-time processes), MARIO, Dem@Care (data integration and health care), Facts4Workers (benefits of semantics in the industry)
ENeL COST: Cooperation already established with LIDER. To be extended with cooperation with FREME (new project), especially on publishing lexicographic data in RDF in the LOD (also with the partner iMinds) The topic of quality checks was here also discussed (with Christophe Lange, Bonn).
ENeL COST: W i t h t h e W 3 C C o m m u n i t y G r o u p „Ontolex“ (supported by LIDER), we discussed the possible representation of Gender issues in the ontolex model (and extensions), as well as the indication of the various temporal information incuded in lexicographic resources. With MixedEmotions we plan a cooperation on the LOD/RDF representation of sentiment words in the authorative dictionaries in the ENeL network
EU Project Networking Session 2015
– Discussion Minutes
Table 4: Industry & Business
• InteresNng contact with STI / BYTE on sharing knowledge for Big Data Europe
• Temporal context has great value in analyzing Big Data. A feedback loop which updates a limited, but easily accessible, LD model gives similarity to the Lambda Architecture.
• PraNcal use of conNnuous event processing demonstrated by Leon Derczynski
• "We want a new networking session in a few months." -‐-‐ Andrea Maurino
• "We could use Graphical VisualisaNon to summarize models" -‐-‐ Aidan Delaney
• "We need profiles to describe Big Data streams" – abstartup
• Strabon and Marmo`a can be used as temporal database
• Reasoning and summarisaNon is a problem with large datasets. Visual and formal summaries are useful for different cases.
• Temporality is important and a near-‐ubiquiNous challenge, in different forms. Making the best decision as you can is possible in real-‐Nme, but it also useful to be able to rewind and determine the support for a decision that was made in the past. cf. databases Strabon and Marmo`a
• Reasoning for streamed data is tough. It can be useful and so it is useful to process streams using summaries of representaNons, to keep things fast.
• InformaNon is volaNle and the volaNlity of informaNon can indicate reliability, as can its age. However, old or very new asserNons are not necessarily correct. Reliability scoring likely evolves best with a Bayesian model.
• Event Registry is an enNty-‐centric event exploraNon tool, from xLime/xLike, useful to other partners (with links to Pheme). (cf. cola.js)
• TELIOS describes visualisaNon of temporal changes on a map, over Nme, from linked data.
• The WDAqua ITN looks at quesNon answering and is concerned with the reliability of data for providing support in answering, and is interested in visits to and from other insNtuNons working on related topics (e.g. Pheme's veracity challenges).