Creating Knowledge from Interlinked Data

Creating Knowledge out of Interlinked Data

Intelligent Information
Management
Collaborative Project 2010-2014
in Information and Communication Technologies

Project No. 257943
Start Date 01/09/2010

http://lod2.eu
EU-FP7 LOD2 Project Overview . 02.09.2010 . Page 1 http://lod2.eu


The emerging Web of Data achievements and challenges

• Web - a global, distributed platform for data, information and knowledge integration
• exposing, sharing, and connecting pieces of data, information, and knowledge on the
Semantic Web using URIs and RDF
Achievements  Challenges
1. Extension of the Web with 1. Coherence: Relatively few,
a data commons expensively maintained links
(currently amounting 25 2. Quality: partly low quality
July 2007 April 2008 September 2008 Billion facts) data and inconsistencies
2. vibrant, global RTD 3. Performance: Still
community substantial penalties
3. Industrial uptake begins compared to relational
(e.g. BBC, Thomson 4. Data consumption: large-
Reuters, Eli Lilly) scale processing, schema
mapping and data fusion still
4. Emerging governmental
in its infancy
adoption in sight
5. Usability: Missing direct end-
5. Establishing Linked Data user tools and network effect
as a deployment path for
the Semantic Web. These issues are closely related
July 2009 and should ultimately lead to an
ecosystem of interlinked
EU-FP7 LOD2 Project Overview . 02.09.2010 . Page 2 knowledge! http://lod2.eu

LOD2 in a Nutshell
Research focus
• Very large RDF data
management
• Knowledge Enrichment &
Interlinking
• Fusion & Information
Quality
• Adaptive, semantic user
interfaces
Use Cases
• Media & Publishing
• Enterprise Data Webs
• Open Gov Data
Main Result
• Integrated LOD2-Stack
for Linked Data lifecycle
management
Partner
Uni Leipzig, CWI, DERI
Galway, FU Berlin,
Semantic Web Company,
OpenLink, Tenforce,
Exalead, Wolters Kluwer,
OKFN 3


LOD2
EC-funded collabarotive project that aims to utilize the Web as an integration platform for
data and information

Linked Data
Linked Data provides the necessary basic technologies and standards to realize the goal
of LOD2.

Linked Open Data
publicly accessible data which is to be integrated into the web and linked among one
another and with non-public contents such as enterprise intranets

Project Highlights
Open Government Linked Data Initiative
Common European platform publicdata.eu

Leading Web 3.0 technologies are combined in the project in to the coherent LOD2
stack (e.g. DBpedia, Virtuoso, Sindice, Silk)



WP1: Requirements, Design & LOD2 Stack Prototype

Use Case High-Level Abstraction



WP1: Use Case Objectives
Objective of WP8:
Applying Linked Data technologies in an enterprise stack
to support Human Resources-related issues.

ENTERPRISE
APPLICATIONS

(Exalead)

Objective of WP7: MEDIA OPEN Objective of WP9:
Supporting content-related & GOVERNMENT Improving accessibility,
production workflows in the PUBLISHING DATA findability and reusability of
media & publishing industry. Open Government Data.
(Wolters (Open
Kluwer Knowledge
Germany) Foundation)



WP2: Storing & Querying Very Large Knowledge Bases

Goal:
 Enabling large-scale, feature-rich & enterprise-ready Linked Data
management solutions

Database Partners in LOD2:
 CWI - Leading open source analytics RDBMS
 OpenLink - Leading Linked data deployment platform

Technological Excellence:
 Creating and publishing metrics for choosing RDF solutions
 Bringing Column Store Technology for Business Intelligence on RDF
 Ground-breaking database innovations for RDF stores
(Dynamic Query optimization, Adaptive Caching of Joins, Optimized Graph
Processing, Cluster/Cloud scalability)



WP2: Linked Open Data For Real In Your Apps

Business Advantages:
 Enrich your application with (free & rich) Linked Open Data
 RDF store technology has 10x lower deployment costs than relational for ragged data

Technological Flexibility:
 Deliver Schema-Last Flexibility and Inference at Relational Data Warehouse Cost and Performance
 Grow as you go: the LOD2 platform dynamically adapts to your usage patterns and structure of your
data
 Integrate, resolve, align anything: Schema, instance identity

Rich Features for complex Applications:
 Advanced SPARQL and SQL query processing
 SPARQL and SQL Federation
 Full Text, Geospatial, Text Search
 Scale-Out on Clusters, Replication



WP3: Goals

General Goal:
 Creation, improvement, repair of knowledge bases

Focus:
 Very large knowledge bases, diverse knowledge, web data
 Refine existing (Virtuoso Sponger, RDF Views, Triplify, D2R) triplification approaches
 Improve schema of knowledge based on data
 Fix problems in knowledge bases e.g. inconsistencies

Techniques:
 semi-automatic machine learning, ontology debugging, NLP, shallow parsing etc.



WP3: Knowledge Base Improvement Cycle

Mutual Refinement Cycle
(with optional Extraction
phase) Modelling
Problems

Repair

Performance
Problems

Disjoint-
Semi- ness
Structured Definitions
structured
Extraction Enrichment

Un- Linkage
structured Validation



WP3: Task Overview
Provenance-Aware Extraction of Linked Data from Existing Structured Formats
 Relational databases, spreadsheets, CMS, logs, XML documents
 Development of D2R Triplify, Virtuoso Sponger

Provenance-Aware Extraction of Linked Data from Unstructured and Semi-
Structured Sources
 HTML, PDF and Office documents with meta-data, wiki code, plain text
 Development of NLP2RDF and DBpedia

Knowledge Base Schema Enrichment
 Learn axioms in knowledge bases, e.g. disjointness, definitions, super-classes
 Development of ORE and DL-Learner

Knowledge Base Repair
 Fix inconsistencies, modeling problems, reasoning performance problems
 Development of ORE

Web Linkage Validator
 Reports whether knowledge base is suitable to be interlinked with others



WP4: Reuse, Interlinking and Knowledge Fusion (1)

Goal: Provide open-source software components for link generation, schema
mapping, data quality assessment and knowledge fusion.



WP4: Reuse, Interlinking and Knowledge Fusion (2)

Technological Excellence:
 Ease the creation of RDF links by using machine-learning as well as link quality assessment
workbench
 Provide for the flexible integration of Web data based on mappings discovered on the Web
 Provide for assessing the quality of Web data and fusing high-quality data.

Expected Outcomes:
 Link discovery tools, linking assist and workbench
 Framework for publishing and discovering expressive mappings on the Web
 Data quality assessment framework providing for a wide range of different quality assessment
policies
 Data fusion components providing various conflict resolution strategies



WP 5: Linked Data Visualization, Browsing and Authoring (1)

WP5 aims to build on and go beyond existing approaches for realizing adaptive Web
user interfaces by:

 Automatic content adaptation

A subset of the domain knowledge is identified, possibly through some reasoning
mechanism, as relevant to the current user and context.

 Adaptive Browsing

Faceted spatial semantic browsing: reusable component for browsing spatial content
in a faceted way (also for mobiles).



WP5: Linked Data Visualization, Browsing and Authoring (2)

Adaptive Semantic Authoring

 semantic widget interface: allows the creation of small reusable interface components for domain-
specific
user interfaces (also for mobiles).

 adaptive widget choreography: enables the automatic generation of user interfaces.

 social networking interfaces: enable users to subscribe to arbitrary information adhering to certain
semantically defined filter criteria.



WP 5: Technologies & Methods (1)

 Semantic pipes:

An engine and graphical
environment for general Web
Data transformations and Mashup.

 Sig.ma:

A service and an end-user
application to access the
Web of Data as an integrated
information space.



WP 5: Technologies & Methods (2)

Site Services: Site Search and Site Widgets
 Widgets (right) provide relevant information,
from Sindice, about the topic of the site.

 Site search (below) provides a rich faceted-browsing
functionality of the site’s widgets.



WP6: Interfaces, Integration & LOD2 Stack (1)

This work package deploys the LOD2 stack, based on the requirements and
prerequisites
defined in work package 1.

The LOD2 stack will be made available as downloadable packages.

While leading work package 6, the following will be delivered:

 Integrated user-interface components
 Integrated LOD2 Stack API components
 Evaluation, Documentation, Tutorials



WP6: Interfaces, Integration & LOD2 Stack (2)

Output WP1

Use Case WP7:
Media & Media &
Publishing Publishing
Yearly releases

Requirements/
Applied on
Prerequisites

Use Case Generates
Open WP8:
Requirements/ Applied
Enterprise Prerequisites packages with Source on Media &
integrated tools
Data web Package Publishing
SAF LOD2 stack
(Software Assembly Factory)
Requirements/ Released on
Starts on 09/2011 • 09/2012
Prerequisites Applied on
• 09/2013
• 09/2014
Use Case WP9:
Government Media &
Data Publishing



WP7: LOD2 for Publishers

Wolters Kluwer Deutschland (WKD):
“Semantic Technologies and Standards are an enabler for the media and publishing
industry to create added-value for their customers with reasonable costs.“

WKD Legal & Regulatory
Companies/Brands Products (Examples)
- Carl Heymanns Verlag - IP, Administrative Law WKD is part of Wolters Kluwer B.V.
- Luchterhand - Civil, Family, Labor Law
- Werner Verlag - Construction Law Customer orientation Worldwide reach
- Carl Link - Publications for Schools/KiTas - Lawyers - Europe
- CW Haarfeld - Public Health Insurance - Tax Accountants - North America
- Deutscher - Magazin „Personalwirtschaft“ - Corporations and SMEs - Asia/Pacific
Wirtschaftsdienst (HR Management) - Fincancial institutions
- AnNoText - SW for Lawyers and Notaries - Health Providers Economic success
- Trigon Data - Public Sector - Revenue 2009 EUR 3,4 bln.
- 18.000 Employees
WKD Tax & Accounting - Listed Amsterdam SE

Companies/Brands Products (Examples)
- Akademische Arbeits- - Tax SW for Consumers
gemeinschaft Verlag
- Addison Group - SW for Tax Accountants
- Schleupen Tax - SW for SMEs with focus
- Wago Curadata Controlling and Accounting



WP7: WKD as a Consumer of LOD Data

Content Supply
Chain of
Content Composing Publishing Customer
Wolters Kluwer Editing Sales Customer
Acquisition Bundling Interfacing Service
Deutschland

(WKD)

Content Acquisition Content Enrichment Enterprise Applications

Acquisition of LOD governmental data Enrichment of WKD data Data integration in Enterprise and other
Costumer Applications
- Laws & Regulations - Enrichment with additional metadata
from the LOD cloud - Integration of customer and WKD data
- Court cases
with data from the LOD cloud
- Automatic Interlinking within WKD data,
- Administrative Rulings
but also into the LOD cloud - Development of new services, e.g.
- Statistical information around metadata economics
Based on:
Based on: Based on:
- Adequate delivery format
- Adequate delivery format - Adequate functionality
- Adequate metadata
- Adequate metadata - Adequate APIs
- Adequate functionality
- Adequate Licensing and IPR - Adequate Licensing and IPR
- Adequate Licensing and IPR



WP7: WKD as a Publisher of LOD Data

Content Supply
Chain of
Content Composing Publishing Customer
Wolters Kluwer Editing Sales Customer
Acquisition Bundling Interfacing Service
Deutschland

(WKD)

Cloud - Publishing Marketing measures

Development of WKpedia Integration in overall marketing
strategy of WKD
- Publishing of enriched governmental
information - Dissemination of LOD2 in media and
publishing sector
- Publishing of legal domain thesauri
- Launching surveys
- Motivating contextualisation in LOD
cloud - Permanent information of customers
Based on: - Sponsoring of conferences
- Adequate functionality Based on:
-Adequate APIs - Clear scope of LOD2 project to support
future publishing paradigms
- Adequate Licensing and IPR



WP8: Towards Linked Enterprise Data Webs (1)
Linked Enterprise Intra Data Webs can fill the gap between Intra-/Extranets and ERP
systems
Facilitates data integration along value-chains within and across enterprises
The pragmatic, incremental, vocabulary based Linked Data approach reduces data
integration costs significantly
Objectives: Web
publishing

 Promote openness and standards
in enterprise data workflows
and applications



WP8: Linked Enterprise Data Use Case Scenario (2)

Wage policy EBI:
 Build an application for surveying wage policy in a company, domain, sector, region, etc.

Scenarios:
 A company wants to know if its wage policy is consistent with the market (in similar and related
companies and sectors).

 A job applicant would like to have an idea about his wage expectations according to his
expertise, profile
and education background

 A governmental agency would like to survey the salaries in a particular region according to an
economic
branch and other parameters



WP8: Linked Enterprise Data (3)

Targeted service:

 A Saas service with different levels of subscription

 The service is a mashup of payroll and HR data of enterprises subscribing to the
service to
build an index store of data facts about wages.

 Different consolidation parameters and key performance indicators (KPI) will be
studied to
provide relevant reports and visualisation interfaces.

 Integration of external datasets in a particular survey: public datasets in the web cloud
or
private datasets of participating companies.

 Privacy issues management: make private and nominative data anonymous.


WP8: Linked Enterprise Data (4)
Preliminary overview:

Search and
EBI
Employees
interface
Full and HR
text database
index

Indexer

Data
Data Data Data Data LODification Taxonomies
consolidatio enrichment, Cleaning and crawler and
n annotation uniformisatio anonymisation
n

RDF
store SPARQL Payroll software
endpoint


WP9: Open Government Data Use Case

publicdata.eu - find and reuse datasets from local, regional and national public
bodies across Europe from a single place



WP9: Who is this for?

 Data literate citizenry
 Data journalists
 Policy experts
 Decision makers
 Mobile and web developers
 Academics / researchers
 Public bodies
 Companies
 Civic society / NGOs
 And so on...



WP9: What will it involve?

 Enabling exchange of metadata between different data catalogues

 Aggregating datasets from existing data catalogues

 Creating a European community of reusers to improve metadata

 Creating mechanisms for capturing derived / related datasets

 Bridge language and topical gaps to associate related information
from all Member States



WP10: Training, Dissemination, Community Building & Fertilization

The general aim of this work package is to establish a worldwide focal point for
academic and industry parties interested in contributing to or taking advantage of
the novel Linked Data methodologies and components, which will emerge in the
project.

In particular, our activities will be targeted at:

 informing the community of the state-of-the-art developments taking place in the
field,

 disseminating the project results in order to foster community building and to
create an
impact on industry and research in Europe and worldwide,

 providing training to interested audiences in the technologies developed
throughout the
project


WP10: Tasks & Timeline
Task 10.1 Training
(M 14 – M37)

 Internal face-to-face training
 External training
 PhD programme

Task 10.2 Dissemination, Community Building & Cross-Fertilization
(M1 – 48)

 Scientific dissemination
 Industrial dissemination
 Online marketing activities across all identified target groups



WP10: LOD2 Dissemination Resources

Website: http://lod2.eu

Weblog: http://lod2.eu/BlogPost

Twitter: http://twitter.com/lod2project

SlideShare: http://www.slideshare.net/lod2project

PUBLINK: http://lod2.eu/Article/Publink.html

Remark: please use #lod2 on twitter for your posts & connect with account: lod2project
many thanks in advance!!



PubLink – LOD2’s Linked Open Data Starter Service
• PubLink helps selected organizations with a focused consulting effort of 10-15 days
to publish and make use out of Linked Data

• PubLink helps to evaluate the LOD2 technologies and to increase the wealth of
Linked Data

• Yearly application deadline in Winter

• 2011 PubLink participants include:

1. Umweltbundesamt GmbH, Austria
2. Greater London Authority
3. Deutsch Bibliographie, Historische Kommission
4. The Parliament of Finland
5. City of Vienna
6. Instituto Canario de Estadística (ISTAC)

• See: http://lod2.eu/Article/Publink.html



WP11: Exploitation and Standardization (1)

Objectives:
 Realizing the vision of the LOD2 project and use case studies
 Standardisation of LOD2 architecture
 Exploitation of knowledge and technical results

Exploitation:
 Use case studies and the industrial and end-user community partners will drive the exploitation.
 Tracking important technical and commercial in information retrieval, data management including
news
and media.
 Publish exploitation plan identifying opportunities, benefits and impact of LOD2 consortium.



WP11: Exploitation and Standardization (2)

Interlectual Property Rights (IPR):
 Core component of the LOD2 stack will be published under open-source license.
 Domain adoptions of LOD2 stack considered on case-by-case basis to protect IPR.
 Strategy ensures that all components of LOD2 are royalty-free.

Standardization:
 Actively participating in appropriate standards bodies.
 Establishing a W3C Linked data interest group.

Orchestration with other projects:
 Encourage take-up of LOD2 technologies by other projects.
 Foster input from other EU projects relevant for the development of LOD2



WP12: Fact Sheet

Project Means
 Instrument: Large-scale Integrating Project
 Objective: Intelligent Information Management  Total Budget: 8,58 M€
 Call: FP7-ICT-2009-5  Total Funding: 6,45 M€
 Duration: 09/2010 – 08/2014  Total Resources: 844 PM

Consortium
 Universität Leipzig (Coordinator)
 Centrum Wiskunde & Informatica
 National University of Ireland in Galway
 Freie Universität Berlin
 OpenLink Software
 Semantic Web Company
 TenForce
 Exalead
 Wolters Kluwer Deutschland 10 Partners from 7 European
 Open Knowledge Foundation Countries



Contact

Address Coordinator

University of Leipzig Dr Sören Auer
Faculty of Mathematics and Computer Science Scientific Project Leader
Institute of Computer Science Phone:+49 (341) 97-32367
Department of Business Information Systems
Fax: +49 (341) 97-32329
Postfach 100920 Email: auer@uni-leipzig.de
04009 Leipzig http://www.informatik.uni-leipzig.de/~auer
Germany
Nadine Jänicke
Project Manager
Phone:+49 (341) 97-32310
Fax: +49 (341) 97-32329
Email: jaenicke@uni-leipzig.de
http://bis.informatik.uni-leipzig.de/NadineJaenicke

Thanks for your attention! http://lod2.eu
LOD2 Title . 02.09.2010 . Page 37 http://lod2.eu

Creating Knowledge from Interlinked Data

Recommended

Recommended

More Related Content

What's hot

What's hot (18)

Viewers also liked

Viewers also liked (19)

Similar to Creating Knowledge from Interlinked Data

Similar to Creating Knowledge from Interlinked Data (20)

More from LOD2 Creating Knowledge out of Interlinked Data

More from LOD2 Creating Knowledge out of Interlinked Data (20)

Recently uploaded

Recently uploaded (20)

Creating Knowledge from Interlinked Data

Editor's Notes