6th DBpedia Community Meeting in The Hague 2016

Marco Brattinga and Arjen Santema, Land Registry and Mapping Agency (Kadaster)

The Dutch Cadastre exploits several national key registers (cadastral registration, topographic map) and information nodes (addresses and buildings, real estate value, cables and pipelines, large scale topographic map). Marco Brattinga is principal consultant at Ordina. Arjen Santema works as consultant tactical information management and innovations at the Cadastre. They share the passion for semantics and practices for publishing data and metadata on the web. Together they developed a framework to describe the data and metadata in a registration in relation to a concept schema that describes what the registration is about and helps to understand this. By separating the concept schema from the data and metadata model they created a taxonomy with domain experts, that contains implementation independent definitions. On the other side they built a data model or an ontology for specific goals. They will present the ideas behind this framework and show some examples from the cadastral registration, the topographic map and the information node addresses and buildings.

Antoine Isaac and Hugo Manguinhas, Europeana

Antoine Isaac and Hugo Manguinhas are members of the R&D team at Europeana.eu, the platform for Europe’s digital cultural   heritage from libraries, museums and archives. We facilitate, coordinate and promote technological innovation for data aggregation, enhancement and dissemination of digital cultural heritage data and its associated services within the Europeana network. Our activities focus notably on data exchange, data quality, multilingualism and search.

Marco de Niet, DEN Foundation

Marco de Niet is the director of the DEN Foundation, the Dutch knowledge centre for digital heritage. He is actively involved in both national and international networks that focus on innovation with cultural heritage assets, including the Europeana Members Council. He is advisor to the policy officers of the Dutch Ministry of Culture that are responsible for the digital heritage strategies on the national level. He is responsible for the ENUMERATE Framework (currently part of Europeana) used for measuring the progress of digitisation in Europe. He is a member of the Dutch Unesco Memory of the World Committee, a board member of the Dutch Museum Register and a member of the Council for Dutch Language and Literature, which enhances the collaboration between the Netherlands and Flanders. He has a background in cultural information science.

17:00

Registration* and gathering on the 10^th floor

* For security reasons at the TNO building in The Hague visitors have to register themselves at the reception desk with their ID or passport. Please make sure that you bring yours!  

17:30

• Introduction to DBpedia/DBpedia Association by Sebastian Hellmann, AKSW/KILT
• General introduction about TNO Data Science department by Jean-Louis-Roso
• Introduction about Platform Linked Data Netherlands (PiLOD) by Erwin Folmer

18:00-19:30

Social event and poster/demo reception with the following projects:

• The Smart Appliances REFerence ontology (SAREF) and Standardization in IoT
• Linked Data in Horticulture
• Semantic technologies in Logistics
• Semantic search in Image Retrieval
• Dark Web
• Smart reasoning for well being

Friday, 12. February:

DBpedia Community Meeting

10:00 - 10:30

Meet & Greet

10:30 - 12:00

Opening Session / chair: Gerard Kuys

room: Auditorium

• (5) Opening by Menno Rasch, Director of KB operations

• (10+5) DBpedia Association Update by Sebastian Hellmann, AKSW/KILT - slides

• (10+2) Digital Heritage in the Netherlands by Marco de Niet, DEN Foundation: Because this is what DBpedia is about as far as Dutch cultural institutions are concerned: in order to create common reference points, a common information infrastructure for heritage information is needed. What is more, public institutions in the field of culture have a responsibility in upholding such an infrastructure. Since the ultimate measure of those institutions' usefulness is in the ease with which a user of heritage information can find the answers to his or her questions, the requirement presents itself that cultural institutions cooperate ever more closely in order to be able to offer cultural information as a  single, interlinked package. In the end, the cultural institutions will find the information on and from their collections to be tightly knit together with the public data as being published by Europeana and the Network Digital Heritage. The presentation is available here.

• (20+5) Keynote #1: by Marco Brattinga and Arjen Santema, Land Registry and Mapping Agency (Kadaster): Marco Brattinga and Arjen Santema talking about the best pratices they developed for the Dutch Land Registry and Mapping Agency (Kadaster). Together they developed a framework to describe the data and metadata in a registration in relation to a concept schema that describes what the registration is about and helps to understand this. By separating the concept schema from the data and metadata model they created a taxonomy with domain experts, that contains implementation independent definitions. On the other side they built a data model or an ontology for specific goals. They will present the ideas behind this framework and show some examples from the cadastral registration, the topographic map and the information node addresses and buildings.

• (20+5) Keynote #2: by Paul Groth, Elsevier: Knowledge Graph Construction and the Role of DBPedia

  In this talk, I want to present some of our recent work on Knowledge Graph construction and the role of DBPedia and other Wikipedia based knowledge in that work. I discuss how having a an updated publicly available knowledge graph is crucial in acting as a reference for constructing internal knowledge graphs. The presentation is available here.

• Introduction to the program by Gerard Kuys, Dutch DBpedia

12:00 - 12:45

Lunch Break

12:45 - 14:15

DBpedia Showcase session

room: Auditorium

• (5) DBpedia+ Data Stack 2015-10 - Release by Markus Freudenberg, AKSW/KILT

• (15) Enriching Cultural Heritage Data with DBpedia by Antoine Isaac, Europeana: For Europeana, getting richer metadata is a priority. It improves access to the nearly 50 million cultural heritage objects, notably by enabling better multilingual retrieval and creating relations between objects. The Europeana Data Model (EDM) allows the ingestion of semantic and multilingual metadata. It notably supports the representation of contextual links to concepts described in third-party data sources such as DBpedia. These rich entities are either provided by data providers in their metadata or selected by Europeana using semantic automatic enrichment. To further enhance its data and improving documents retrieval across languages, Europeana is now working on a “semantic entity database”. DBpedia will be one of the source datasets used as an anchor for more domain specific vocabularies. Find the compelte presentation here.

• (10) DBpedia Wayback Machine by Patrik Schneider, Siemens and WU Wien: DBpedia is one of the biggest and most important focal point of the Linked Open Data movement. Despite its multiple services, it lacks a fine-grained wayback mechanism to retrieve historical versions of resources at a given timestamp in the past, thus preventing systems to work on the full history of RDF documents. In this talk, we present (a) the framework that serves this mechanism and is publicly offered through a Web UI and a RESTful API, following the Linked Open Data principles; and (b) the usage of the Wayback Machine in the CityDataPipline. Find the complete presentation here.

• (10) BlueSky - Knowledge Diviner - DBpedia demo by Richard Nagelmaeker: A new view on data interaction as a reference point for data-driven architecture. Find the complete presentation here.

• (5) GOOSE by Laura Daniele, TNO Find the complete presentation here.

• (5) DBlexipedia: A nucleus for a multilingual lexical Semantic Web by Christina Unger, CITEC: Natural language-based applications using DBpedia face the challenge that they require knowledge about how the ontology elements are verbalized in natural language. In order to provide such knowledge at the required scale and thereby leverage the use of DBpedia in different applications, we construct a lexicon for the DBpedia ontology by means of existing automatic methods for lexicon induction. It contains 11,998 lexical entries for 574 different properties in three languages: English, German, and Spanish. Just like DBpedia provides a hub for Semantic Web datasets, this lexicon can provide a hub for the lexical Semantic Web, an ecosystem in which lexical information are published, linked, and re-used across applications. Find the complete presentation here.

• (5) DBpedia Historic data by Raphael Boyer, DBpedia FR / INRIA Find the complete presentation here.

• (5) Using Elasticsearch + DBpedia to maintain a searchable database of global power plants. by Chris Davis: Many countries are undergoing an energy transition, although relevant data about their portfolio of power plants is often only found as a mix of official and crowdsourced data.  To navigate this, we have set up an Elasticsearch instance containing this data (http://enipedia.tudelft.nl/Elasticsearch.html).  To help synchronize the collection of the 4000+ Wikipedia articles on global power plants with their latest edits, we use a single SPARQL query to DBpedia that performs a category traversal to retrieve information on the latest revision IDs of the articles (https://github.com/cbdavis/wikipedia-power-plants). Find the complete presentation here.

• (5) Linked Data Reactor by Ali Khalili: The LD-R framework combines several state-of-the-art Web technologies to realize the vision of Linked Data components. LD-R is centered around Facebook's ReactJS and Flux architecture for developing Web components with single directional data flow. LD-R offers the first Isomorphic Semantic Web application (i.e. using the same code for both client and server side) by dehydrating/rehydrating states between the client and server.

• (10) Using DBPedia in Europeana Food and Drink by Vladimir Alexiev / Ontotext: The Europeana Food and Drink project collects cultural heritage objects for and develops applications related to Food and Drink heritage. As part of the project, Ontotext developed a FD Classification based on Wikipedia/DBpedia Categories, a semantic enrichment service that annotates each CHO with FD Topics and Places, and a semantic application that implements hierarchical semantic facets and semantic search for these facets. We'll also be packaging the enrichment as a service for others to use in a crowdsourced annotation application.  We will explain how we used Categories to build a domain-specific gazetteer, used external datasets (eg UMBEL domains and DBTax types), correlated DBpedia places to Geonames to use the place hierarchy, and the workings of the semantic application. Find the complete presentation here.

• (5) FREME by Nilesh Chakraborty. Open Framework of e-Services for Multilingual and Semantic Enrichment of Digital Content. Find the complete presentation here.

• (5) Using DBpedia for improved Vaccine Information Systems by Monika Solanki, University of Oxford Find the complete presentation here.

14:15 - 14:30

Coffee break (15”)

PS1: DBpedia ontology / chair: Monika Solanki, University of Oxford 

Room: B

• DBpedia ontology survey results & discussion

• Agenda:

  1. Introduction to the DBpedia ontology

  2. talks from people who are extensively and aggressively using the ontology:

  • (7) WebProtégé demo & aspect oriented programmingby Ralph Schäfermeier & Alexandru Todor / FU Berlin, 

  • (5) classification ontology by Gerard Kuys / Ordina

  • (5) DBpedia mappings quality problems, Vladimir Alexiev / Ontotext

3. DBpedia ontology survey results

4. Proposed plan to address action points from the survey results

5. Discussion

PS2: DBpedia & Heritage: Challenges and opportunities of reference data for digital heritage / chair: Enno Meijers, Dutch DBpedia (Find the complete presentation here.)

Room: Auditorium

• (15) Building an ecosystem of networked references by Hugo Manguinhas, Europeana:  Over the past five years, the amount of contextual entities in Europeana’s metadata has grown considerably. These entities are provided as references as part of the metadata delivered by Europeana or selected by Europeana semantic automatic enrichment. Pursuing their efforts towards the creation of a semantic network around cultural heritage objects, Europeana and its partners providers and aggregators are investigating ways to better exchange vocabulary data and manage co-references/alignments between vocabularies. In this presentation we will explore the potential of tools such as OpenSkos and Cultuurlink for supporting the building of networked references. Find the complete presentation here.

• (15) RML - generating high quality Linked Data by Anastasia Dimou, iMinds: Despite the significant number of existing tools, incorporating data from multiple sources and different formats into the Linked Open Data cloud remains complicated. The RML tool chain developed by iMinds provides a generic solution, based on an extension over R2RML, for mapping data in a source-agnostic and extensible way, while facilitating the definition of mappings of multiple heterogeneous sources. Find the complete presentation here.

• (15) Histograph: geocoding places of the past by Joop Vanderheiden, RCE:  Histograph: a historical geocoder for search and standardization of place names throughout history. Histograph collects and links place names and uses these to georeference and standardize place names in time; currently, sources used include birth places of Dutch East India Company crew members, monastry records and historical census data. Find the complete presentation here.

• (15) "Illegal newspapers in the WOII" Wikipedia/DBpedia project by Olaf Janssen, KB: Olaf Jansen, the wikipedian-in-residence at the KB works on describing nearly 1300 illegal newspapers printed in the WOII in the Dutch Wikipedia. The fulltext of the newspaper will be available at the Delpher service of the KB. The newspaper articles will be linked to relevant persons and places using the DBpedia. Find the complete presentation here.

15:45 - 16:00

Coffee break (15”)

16:00 - 17:15

BS1: DBpedia Dev session / chair: Dimitris Kontokostas

Room: A

Session for developers to talk about technical issues and challenges in DBpedia including:

• (5+5) Recent Quality improvements in DBpedia by Dimitris Kontokostas / AKSW/KILT

• (7+5) Mappings wiki with a Git-based approach by Alexandru Todor / FU Berlin

• Hosting

• DBpedia+ (how to contribute and interlink data)

• mini roadmap

BS2: DBpedia Tutorial by Markus Freudenberg, AKSW/KILT

Room: B

a one hour tutorial about Linked Data and DBpedia

The tutorial will start with an open talk to assess the level of the audience and then either start with a general introduction to Linked Data or go more into detail and provide an overview and tips and tricks on DBpedia components and  what we can do with it.

BS3: DBpedia & NLP (partially focused on Cultural Heritage) chair: Sebastian Hellmann, AKSW/KILT

Room: Auditorium

In this session we will investigate the application of Linked Data on Language Technologies, especially entity linking. The domain is focused partially on Digital Humanities

(20) Results of the LIDER project by Christina Unger, CITEC: One of the outcomes of the LIDER support action are guidelines to facilitate the discovery, reuse and exploitation of existing linguistic resources, aiming at the establishment of a new Linked Open Data (LOD) based ecosystem of free, interlinked, and semantically interoperable language and media resources (corpora, dictionaries, lexical and syntactic metadata, images, etc.) for multilingual, cross-media content analytics. In order to show how these guidelines can be realized, I will showcase example services for discovering and querying relevant linguistic resources and for using and linking LOD-aware NLP services. The full presentation is available here.

(20) TellMeFirst ­ A Knowledge Domain Discovery Framework by Giuseppe Futia: TellMeFirst (TMF) is an open­-source framework that leverages the DBpedia knowledge base and the Wikipedia corpus for classifying documents, achieved by computing a similarity score between the target document and an initial training set. Each DBpedia entity identifies a document of the training set that includes all Wikipedia paragraphs in which the entity appears as wikilink. The training set is composed by entities covered in DBpedia from different knowledge domains (such as Cultural Heritage, Politics, History, Science, Sport). Nevertheless, cultural institutes, companies, and public administrations are much more interested in a classification system that exploits only a subset of these entities, specific for their purposes and needs. During the talk, we will introduce both the transformation pipeline (based on DBpedia Spotlight project) for building a general-­purpose training set and a configurable process for building a domain training set. Then, we will report on the differences between classification results obtained by evaluating TMF with the two training sets in an example scenario. The full presentation is available here.

(20) Mapping the Bio-economy using DBpedia Spotlight by Chris Davis:
A large variety of industrial processes use petroleum-derived feedstocks, and there is significant discussion about creating a bio-based economy which replaces these with more renewable resources.  So what is the bio-economy and which feedstocks are people using, and for which purposes?  To answer this, we collected 78,000 abstracts from journal articles, processed them using DBpedia Spotlight, and created matrices showing which organisms are co-mentioned in abstracts with particular technologies and applications. The full presentation is available here.

17:15 - 18:00

Closing session & networking

18:00 +

Continue networking in a near pub over beer