8th ENDORSE follow-up event:
Merging terminology and knowledge: the power of EuroVoc and IATE through semantic integration
Location: Online webinar
Date: 18 April 2024
Time: 10.00-11.30 CETConnection information:
Meeting link:
https://ecconf.webex.com/ecconf/j.php?MTID=mef86d3523546c03ffc8d06e7207464c1
Meeting number:
2744 563 6075
Password:
u6DQrcPm@35 (86377276 from phones)
Speakers:
Jochen Hummel, CEO, Coreon GmbH
Michael Wetzel, MD, Coreon GmbH
Documents and media
Abstract
We outline the potential of merging a knowledge system (Eurovoc) with a terminology resource (IATE).
Eurovoc, 7000+ nodes, semantically clearly structured top-down, on the one hand and IATE, 600,000+ concepts, terminologically very rich on the other hand, are overlapping yet complementary resources. Let's imagine these two resources woven into one seamless system.
We discuss how to merge, how it looks and works, interoperability and formats, filtering and querying.
We illustrate what its direct benefits are for both Eurovoc as well as IATE maintenance (e.g. quality control, identification of gaps as well as doublettes) and usage.
We spotlight a couple of use cases what such a resource - a linguistically rich knowledge graph - delivers as value, such as: a) text annotation: leveraging all the conceptual and terminological details from IATE and b) auto-correct machine translations produced by eTranslation using GenAI, leveraging the contextual indicators from Eurovoc to address homonymy, ambiguity.
Having shared our knowledge and experiences, we then invite participants to discuss how an MKS - Multilingual Knowledge System - enables multilingual AI and to nail down next steps for enabling development and maintenance of multilingual reference data.
7th ENDORSE follow-up event:
Exploring the potential of Linguistic Linked Data in the LLM era
Date: 28 February 2024 10.00-11.00 (CET)
Speaker:
Patricia Martín Chozas
Ontology Engineering Group, Universidad Politécnica de Madrid
Connection information:
Meeting link:
https://ecconf.webex.com/ecconf/j.php?MTID=mdeeb8ca31c8dbdfcc7f3609f945efd14
Meeting number:
2742 484 6021
Password:
MPtpbCC$866 (67872220 from phones)
Abstract
In recent years, many research efforts focused on the generation of Linguistic Linked (Open) Data (LLOD), driven by their potential to enhance Natural Language Processing (NLP) tasks. Researchers across the globe have dedicated substantial efforts to curate, standardise, interconnect, and leverage language resources, pursuing the generation of Linguistic Linked Data and the population of the LLOD cloud. However, the rise of Large Language Models (LLMs) has challenged conventional NLP methodologies, as they are able to better capture complex linguistic and semantic patterns as they employ deep learning architectures trained on vast amounts of textual data. Therefore, this talks presents three main objectives: first, to give an overview of the LLOD cloud status, including latest advances in LLOD-aware technologies and representation models; second,to identify shortcomings of current LLMs in different NLP tasks, specially, domain oriented; third, to explore the possibilities of incorporating LLOD into LLMs, both for data pre and postprocessing and for knowledge injection into the LLM architecture. Since the latter is still an unexplored territory, the talk intends to identify potential research areas open for discussion.
6th ENDORSE follow-up event:
The European Legislation Identifier (ELI and ELI-DL) – Use cases and implementations
Date: 19 October 2022 10.00-12.00 (CET)
Speakers:
John Dann (Director of the Official Journal of Luxembourg, Ministry of State, Luxembourg)
Jean Delahousse (Semantic web and knowledge graph consultant, ELI Taskforce expert)
María Represa Martín (Deputy director of the Documentation Department, Official State Gazette National Agency, Spain)
Thomas Francart (Knowledge engineering consultant, ELI Taskforce technical expert)
Christie Damnet (Data and process analyst), Laura Liudvinavičiūtė (ICT application officer) (European Parliament)
The recording is available here.
The slides are available:
Abstract:
Nowadays legislation is widely available online and accessible in various digital formats. However, the way in which legal information is organised, varies across the different legal systems of EU Member States, a fact that tends to hinder the discovery, exchange and reuse of this information.
The European Legislation Identifier (ELI) is a solution which aims to address this issue by providing a way to uniquely identify national and EU legislation online.
Born in the framework of the Forum of Official Gazettes, the ELI project was officially endorsed by the Council of the European Union. The ELI initiative is being driven forward by a group of EU Member States together with the Publications Office of the EU. A dedicated Taskforce, chaired by Luxembourg, shares expertise and helps EU Member States implement the new identifier.
The webinar will start with an introduction by John Dann, Director of the Official Journal of Luxembourg at the Ministry of State in charge of publishing and disseminating Luxembourg‘s legislation through the legilux.public.lu portal. He is the founding father of the standard “European Legislation Identifier” (ELI), chair of the “ELI Taskforce” of the Council of the EU and will present how the project started, the aim it addresses and provide a brief introduction of the different elements of ELI, as well as the use case for Luxembourg.
Jean Delahousse, semantic web and knowledge graph consultant and ELI Taskforce expert will highlight the details of the ELI project, in particular the open ontology/data model for Official Gazettes exposing descriptions of official legislation, as a backbone to build a coordinated and interoperable Europe in the legal domain.
María Represa Martín, deputy director of the Documentation Department, Official State Gazette National Agency in Spain, will present the use case of ELI for Spain. Spain has been assigning ELIs to its regulations since 2018. In a decentralized country, this implementation has not been limited to the national level, but rather it has reached the regional and even local level. An experience that has made it possible to lay the foundations for interconnecting a great heterogeneity of legal information systems, through the implementation of a unique URI template, a set of minimum common metadata and controlled vocabulary tables. A use case that speaks about challenges, interconnection, interoperability and a lot of coordination.
Thomas Francart, knowledge engineering consultant and ELI Taskforce technical expert, will present ELI-DL and will provide a walkthrough of the ELI and ELI-DL model to describe legal Documents and Activities.
Christie Damnet, Data and process analyst, and Laura Liudvinavičiūtė, ICT application officer, both from the European Parliament, will present the use case of ELI for draft legislation (ELI-DL) for the European Parliament, applying the ELI-DL ontology as data model in the scope of legislative applications.
5th ENDORSE follow-up event:
Data FAIRification at Scale – Semantic Core Capabilities and Implementation
Date: 31 May 2022 10.00-11.30 (CET)
Speaker:
Martin Romacker (Senior Principal Scientist, Data & Analytics, F. Hoffmann-La Roche, Basel Switzerland)
The recording is available here.
The slides are available:
Abstract:
The pharmaceutical industry has started to massively invest in digital transformation to become data-driven and, thereby, to increase its productivity. The FAIR principles have been adopted as a blueprint to drive our data strategy and to make this transformation successful. However, Data FAIRification at scale has deep implications on the ways we work along our data management value chain affecting people, culture, processes and information technology. Consequently, semantic technologies are almost mandatory to build and represent the necessary core capabilities: terminology management, (meta)data element registries and conceptual modelling. But just building these capabilities is not sufficient, they also have to be adopted in the organization both in terms of data integration and data governance which necessitates a long term commitment and strategy.
At Roche, we have successfully established a semantic backbone to support our digitalization efforts with a high acceptance in the business as we can fully customize and contextualize the reference data (terminologies, metadata models and conceptual models) whilst keeping them fully governed and FAIR at the same time.
The presentation will cover business related aspects and give an introduction to our high level architectural approach. We will then focus on FAIR by design by walking through our capability stack and show we apply our semantic reference data in the various application contexts.
4th ENDORSE follow-up event:
Corporate services on reference data management - VocBench and ShowVoc in practice
Date: 30 March 2022 10.00-12.00 (CET)
Speakers:
Denis Dechandon, Anikó Gerencsér, Mihai Paunescu (Publications Office of the European Union)
Christine Laaboudi (Eurostat, European Commission)
Honza Förster (Cogni.Zone, DG EMPL, European Commission)
The recording is available here.
The slides are available:
Abstract:
In the era of semantic technologies, it is necessary not only to create, update and maintain metadata and reference data, but also to effectively share, visualise and align this data, so as to enhance interoperability. This requires specialised knowledge and established validation procedures, as well as the proper tools and adapted dissemination chains.
The Publications Office of the European Union provides corporate reference data management services for the DGs of the European Commission and similar to other interested stakeholders (EU institutions, agencies and national public administrations). The services focus not only on supporting users in reference data related requests and questions, but also on easing the use of semantic technologies and standards, and supporting the creation and maintenance of reference data and on the other hand on providing the necessary infrastructure (tools and platforms) for maintaining and for disseminating reference data.
In this context a corporate access is offered to all requesting EU institutions, agencies and national public administrations to two semantic applications: VocBench and ShowVoc.
VocBench is a web-based, multilingual, collaborative development platform for managing ontologies, thesauri, lexicons, generic code lists and authority lists. Designed to meet the needs of semantic web and linked data environments, VocBench development has also been driven by the feedback gathered from a community of users made of public organizations, companies and independent users looking for open source solutions for maintaining their reference data and other controlled vocabularies.
On the other hand, ShowVoc is a web-based, multilingual, platform for publishing, browsing and inspecting the same controlled vocabularies as listed above. It further entails cross-dataset features, such as global search and a translation API benefit from the presence of different datasets in order to realize a multilingual resource for term reference and authoritative term translation.
ShowVoc naturally complements VocBench. Initially funded by the European Commission ISA² programme and currently by the Digital Europe Programme (DIGITAL), the development of both applications is managed by the Publications Office of the EU. VocBench and ShowVoc are open source solutions that can be reused and further improved by any interested professionals, while the EU national public administrations maintaining vocabularies in the scope of Legivoc can benefit from a free access to the VocBench and ShowVoc corporate instances made available by the Publications Office.
Additionally, a new tool is being conceived. Currently called “Semantics Elicitation Platform”, it is intended to be used independently or as a companion to VocBench, processing and managing unstructured information (e.g. text, various kind of media) and eliciting knowledge out of it, in the form of structured information modelled according to target ontology vocabularies. The project aims at realising an open environment for automatic sharing and provision of text-engineering components based on the same principles of Linked Open Data (a text mining platform).
The webinar will start with the presentation of the corporate reference data management services offered by the Publications Office, followed by two showcases presented by users from the European Commission (DG ESTAT and DG EMPL) who will speak about their experience of using VocBench and ShowVoc in practice. The third part of the event will provide a short demo of both tools and give the opportunity to participants to ask questions.
3rd ENDORSE 2021 follow-up event:
Wikibase and the EU Knowledge Graph as a use case
16 November 14.00-15.30 (CET)
The video is available here.
The slides are available:
The speakers will be:
Max De Wilde (DG CNECT, EC), Dennis Diefenbach (The QA Company), Roberto Musmeci (DG REGIO, EC) and Georgina Burnett (WMDE)
Abstract:
Wikibase is a free, open source software developed and maintained by Wikimedia Deutschland. Originally developed to run the free knowledge base Wikidata, Wikibase is available as a flexible and collaborative software solution that allows users to set up and maintain their own linked database.
The EU Knowledge Graph, available at The EU Knowledge Graph - EU Knowledge Graph (linkedopendata.eu), is using Wikibase as the underlying infrastructure. Wikibase enables multilingual, collaborative, user friendly and open access to the EU Knowledge Graph which is collecting and integrating structured information related to the European Union.
The largest part of the graph is represented by projects financed by the European Union Cohesion funds. This data is exposed to citizens in Kohesio available at https://kohesio.eu .
In this session, we will give an introduction to Wikibase and present the EU Knowledge graph as a use case.
Second ENDORSE 2021 follow-up event:
Linked Data Platforms in and across EU Member States
On 30 September (14.00 – 17.00, CET) by
Pieter van Everdingen: Platform Linked Data Netherlands, Innovation via meaningful connections
Platform Linked Data Netherlands (PLDN) is an open network community of linked data experts in the Netherlands, who would like to share their knowledge and experiences from the linked data projects they are involved in with everyone who has an interest in linked data and those that would like to get started with their own linked data activities. It has developed into a vibrant community with several hundred active members and a few thousand interested parties.
PLDN activities concentrate around organising networking events, facilitating activities for working groups and releasing publications on linked data topics, like the Gems of Linked Data Applications book (available in Dutch), which contains successful examples of linked data initiatives from our PLDN network that we would like to share with a larger audience. We are also very proud of a very different deliverable that we have created: Play-LOD, which is a card board game that can be used to explain the basic concepts of linked data in an easy to understand and playful way. And to make community members more hands-on with linked data, we stimulate them to experiment with linked data via our free to use PLDN lab environment for which we also organised a number of introduction courses. And we work with organisations in our network to start up new activities and new promising linked data initiatives that can possibly be financed or co-financed from innovation budgets and grants within the EU and the Netherlands. PLDN is completely open for any work co-operation that would lead to further adoption of linked data and more best practices.
Raf Buyle and Veronique Volders: Semantics at your fingertips - How Belgian municipalities create Linked Open data without additional effort
Flanders is building an open ecosystem based on Open Standards and Linked Data. There are about 300 local governments in Flanders, of which over 65% is publishing their local legislation as reusable machine-readable information.
With the use (and reuse) of templates, standards and digitalization of the legal drafting process, linked data allows to reuse more easily across local governments and other stakeholders, predict the impact of decisions before they are made and therefore take optimal decisions, while making information on legislation easily accessible and reusable to anyone.
In this session, we share how we have crossed the chasm: how we introduced our initiative to innovators and early adopters, how we onboarded the early majority. Also, we will share which new Linked Data initiatives Flanders is initiating in the context of the COVID-19 resilience.
Jakub Klímek: Guaranteed and semantically interoperable public government data published and shared as Linked Open DataIn 2021, legislation was adopted in Czechia, introducing the obligation of public administration institutions to provide their public data as open data and, more importantly, to consume the open data whenever it is necessary to access the public data of another institution. Consequently, the Ministry of the Interior of the Czech Republic (MOI-CZ) introduced the so-called Public Data Space (PDS) and is currently working on the necessary infrastructure. It will consist of public data published as open data with additional guarantees ensuring the FAIR principles. The data in PDS will be cataloged in the National Open Data Catalog, along with all the other open data. In addition, the data providers will guarantee the formal correctness and availability of the data to the other institutions. To ensure the interoperability of the shared data, formal open standards (FOSes), as defined in EU DIRECTIVE 2019/1024, will be issued by MOI-CZ for the relevant data domains. A FOS prescribes the allowed data formats and data structures. At its core, it prescribes an RDF representation adhering to the principles of Linked Open Data. Other formats, such as JSON or CSV, can also be introduced if necessary, maintaining transformability to RDF through JSON-LD and CSV on the Web. The semantic interoperability in PDS will be achieved by mapping the data structures defined in FOSes to the public administration ontology, which we call the semantic government vocabulary (SGoV). SGoV defines important domain concepts, relationships between them, semantic mappings among concepts from different domains, and also the ISA Core Vocabularies. The FOSes themselves, including data structure definitions, their expression in different data formats, transformations between the formats, mapping to SGoV, and also documentation, will be generated from SGoV semi-automatically. A set of tools for the management of SGoV will be demonstrated in the talk. We will conclude with the possibility of extending our approach to the whole EU.
First ENDORSE 2021 follow-up event:
Personal Knowledge Graphs: why, what, and where to?
By Ivo Velitchkov, on 2 July (10.00 – 11.30, CET)
The market for Personal Knowledge Management (PKM) software boomed in the last 18 months. As in many other digital areas of accelerated growth, the pandemic served as a catalyst, but the trend was evident before that. And unlike the previous generation of PKM tools, the most successful ones in this wave manage the data as knowledge graphs. Personal Knowledge Graphs (PKG) fundamentally change the way we manage our information, currently trapped in emails, documents, presentations, notebooks and social networks. Many PKG users report huge boost in both productivity and creativity.
This webinar will explore various PKG use cases from intelligent note-taking and project management to learning and research, what kind of capabilities that serve these use cases stabilize in this ecosystem and what could be the impact of PKG on collaborative knowledge management and the adoption of enterprise knowledge graphs.
It seems the role of Personal Knowledge Graphs will only grow in the coming years but the way graphs and applications interact is going to change, transforming the whole ecosystem.
The Publications Office of the European Union and the ISA² programme of the European Commission work together to create an opportunity to link expertise to tackle current and future issues in the fields of interoperability and knowledge management.
Both benefit from their central positions. As a provider and reference organisation for semantic assets, experience, knowledge and technology sharing (standardisation, reference data, semantic assets and technologies, consultancy, guidance and support) for EU and national public services, the Publications Office offers open access for all to a specialised website, EU Vocabularies, and builds on knowledge-sharing platforms such as GitHub and Joinup to promote toolchain building blocks. The ISA² Programme from its side supports the development of digital solutions that enable public administrations, businesses and citizens in Europe to benefit from interoperable cross-border and cross-sector public services.
The EuropeaN Data conference On Reference data and SEmantics
creates an opportunity to link expertise to tackle current and future issues in the fields of interoperability and knowledge management. It brings together a large audience made up of simple users, practitioners, experts, researchers and solution seekers and providers in EU institutions and agencies as well as in national public services and universities.
The very first ENDORSE edition took place from 16 to 19 March 2021. You can view the conference proceedings on the conference website or download a free copy here.
The Publications Office of the European Union and the Interoperability unit of the DIGIT (European Commission) called for the creation of the ENDORSE exchange platform during the closing session of the ENDORSE 2021 event on 19 March 2021.
Our aim is to intensify the exchange of information between simple users, practitioners, experts, researchers and solution seekers and providers in EU institutions and agencies as well as in national public services and universities.
Would you like your voice to be heard and to be part of an active and forward-looking community? Join our community!