Periodic Reporting for period 2 - ELG (European Language Grid)
Período documentado: 2020-07-01 hasta 2022-06-30
The project addressed the fragmentation of the European LT industry by establishing the ELG as the primary platform and marketplace for the European LT community, both industry and research. The ELG has been developed to be a scalable cloud platform, providing, in an easy-to-integrate way, access to thousands of commercial and non-commercial LTs for all European languages, including running tools and services as well as data sets and resources. It enables the European LT community to upload their technologies and data sets into the ELG in an easy and efficient way, to deploy them through the grid, and to connect with other resources. The ELG is boosting the Multilingual Digital Single Market towards a thriving European LT community, also addressing the threat of digital language extinction.
The project was structured into three inter-related areas: the ELG Platform, the ELG Content and the ELG Community.
The project was structured into three inter-related areas: the ELG Platform, the ELG Content and the ELG Community.
Releases 2 and 3 of ELG: The interim release was launched in March 2021 as scheduled, with enriched functionalities focusing on the LT provider side. The final release of the ELG platform was published in March 2022 and integrates all components and functionalities required for a fully operational platform for LT providers and consumers. During the second reporting period, the implementation of the ELG platform has been continued with the development and integration of new components, as planned in the initial workplan, and the improvement and enhancement of the previously integrated components. The design and implementation has taken into account the functional and technical requirements and technological decisions made in the previous reporting period, the continuous feedback from ELG users, pilot and collaborating projects, and the specifications set to address the import of metadata records from other repositories and infrastructures.
Functional Services integrated into the ELG: During the course of the second reporting period, the inventory of tools and services integrated in the ELG plat-form has grown hugely. At the end of the first period (M18) there were 115 fully-integrated LT tool/service metadata records in the ELG platform, by the end of the project this number had grown to over 900, with nearly 200 further services still in the process of integration at this point. However these numbers are an under-estimate of the true coverage of the platform, as one metadata record may offer multiple service functions (e.g. one ser-vice providing both part-of-speech tagging and lemmatisation) and/or may offer support for several different languages (e.g. a multilingual translation model that can translate from any of several source languages into the same target language). Counting these function and language combinations individually gives a total of 1,948 distinct services fully integrated by the end of the project, in addition to a further 2,000+ metadata records repre-senting LT tools or services that are not fully integrated with the ELG platform API specifications.
LRs integrated into the ELG: The second important ingredient in the platform is the population of the ELG catalogue with datasets and language resources. During the second reporting period, the provision of non-functional content through the ELG platform was a high priority as well as the ensuring of an appropriate legal framework for such content and all other data within the platform. Thanks to a huge integration effort, 10,061 metadata records could be integrated. This number of datasets accessible through the ELG platform goes far beyond our expectations.
Open Calls for Pilot Projects: Within the ELG runtime, open calls for small-scale demonstration projects based on the call ICT-29a) specification were specified and implemented, using the Financial Support to Third Parties (FSTP). Two open calls were organised and 15 pilot projects in total were selected for funding and executed, amounting for a funding of EUR 1,948,311 in total. The projects’ results were included in the ELG platform to allow for wide dissemination and testing and external evaluation by other entities and/or the public. All 15 projects were concluded very successfully. The main goal – to attract SMEs and research organisations to either contribute additional tools or resources to the ELG platform or develop applications using Language Technologies available in the ELG platform – was more than achieved.
META-FORUM 2020, 2021 and 2022 – National Competence Centres (NCCs): Within the first six months of the project, the network of 32 NCCs had beeb successfully set up. During the second half of the project, the NCCs were responsible for conducting the ELG workshops which were planned as a dissemination activity on national levels within the European countries. Due to the COVID-19 pandemic, from the 20 workshops that were carried out,16 were held in an online format, one in a hybrid format and three took place on-site.
Because of the pandemic situation, META-FORUM 2020 and 2021 took place as an online conferences over three days with a half-day programme each, containing presentations, panel discussions and large project expos. Because of the improved circumstances of the COVID-19 pandemic in June 2022, META-FORUM 2022 was organised as a hybrid event over two days in Brussels. All three events proved to be very successful and the organisers received a lot of positive feedback from various stakeholders both during and after the events.
Functional Services integrated into the ELG: During the course of the second reporting period, the inventory of tools and services integrated in the ELG plat-form has grown hugely. At the end of the first period (M18) there were 115 fully-integrated LT tool/service metadata records in the ELG platform, by the end of the project this number had grown to over 900, with nearly 200 further services still in the process of integration at this point. However these numbers are an under-estimate of the true coverage of the platform, as one metadata record may offer multiple service functions (e.g. one ser-vice providing both part-of-speech tagging and lemmatisation) and/or may offer support for several different languages (e.g. a multilingual translation model that can translate from any of several source languages into the same target language). Counting these function and language combinations individually gives a total of 1,948 distinct services fully integrated by the end of the project, in addition to a further 2,000+ metadata records repre-senting LT tools or services that are not fully integrated with the ELG platform API specifications.
LRs integrated into the ELG: The second important ingredient in the platform is the population of the ELG catalogue with datasets and language resources. During the second reporting period, the provision of non-functional content through the ELG platform was a high priority as well as the ensuring of an appropriate legal framework for such content and all other data within the platform. Thanks to a huge integration effort, 10,061 metadata records could be integrated. This number of datasets accessible through the ELG platform goes far beyond our expectations.
Open Calls for Pilot Projects: Within the ELG runtime, open calls for small-scale demonstration projects based on the call ICT-29a) specification were specified and implemented, using the Financial Support to Third Parties (FSTP). Two open calls were organised and 15 pilot projects in total were selected for funding and executed, amounting for a funding of EUR 1,948,311 in total. The projects’ results were included in the ELG platform to allow for wide dissemination and testing and external evaluation by other entities and/or the public. All 15 projects were concluded very successfully. The main goal – to attract SMEs and research organisations to either contribute additional tools or resources to the ELG platform or develop applications using Language Technologies available in the ELG platform – was more than achieved.
META-FORUM 2020, 2021 and 2022 – National Competence Centres (NCCs): Within the first six months of the project, the network of 32 NCCs had beeb successfully set up. During the second half of the project, the NCCs were responsible for conducting the ELG workshops which were planned as a dissemination activity on national levels within the European countries. Due to the COVID-19 pandemic, from the 20 workshops that were carried out,16 were held in an online format, one in a hybrid format and three took place on-site.
Because of the pandemic situation, META-FORUM 2020 and 2021 took place as an online conferences over three days with a half-day programme each, containing presentations, panel discussions and large project expos. Because of the improved circumstances of the COVID-19 pandemic in June 2022, META-FORUM 2022 was organised as a hybrid event over two days in Brussels. All three events proved to be very successful and the organisers received a lot of positive feedback from various stakeholders both during and after the events.
The most crucial tangible outcome of the project is the ELG cloud platform itself. The final release was launched to the public in February 2022 (M38) with more than 12,000 metadata records and more than 800 functional services. By the end of M42 it hat approx. 750 registered users. With that, ELG is by a large margin the biggest LT platform globally, with regard to recorded datasets, listed organizations, and registered users.
Both Open Calls for Pilot Projects received an overwhelming responses: In each call more than 100 submitted project proposals were accepted for evaluation (M15). After the evaluations, which resulted in more than 600 individual and 216 summary reports, 15 pilot projects were selected in June 2020 and March 2021. They have started their work in July/August 2020 and February/March 2021 respectively. All projects were completed with great success. The final ELG conference, META-FORUM 2022 in Brussels was very successful, it attracted 100 participants on the spot and up to 300 participants online.
The ELG consortium has compiled a book that documents all aspects of the project and the initiative. The book consists of 30 chapters, the chapters were authored by the responsible work package partners and by the pilot project managers. The book had not been planned as part of the project work plan but was realized in addition to it. The ELG project coordinator is in the process of founding an ELG legal entity that will maintain the ELG infrastructure in the future. For this a transfer document is being prepared in which the details of the transition from ELG as a project to ELG as a registered association are regulated. The association is expected to be in place by the end of 2022.
Both Open Calls for Pilot Projects received an overwhelming responses: In each call more than 100 submitted project proposals were accepted for evaluation (M15). After the evaluations, which resulted in more than 600 individual and 216 summary reports, 15 pilot projects were selected in June 2020 and March 2021. They have started their work in July/August 2020 and February/March 2021 respectively. All projects were completed with great success. The final ELG conference, META-FORUM 2022 in Brussels was very successful, it attracted 100 participants on the spot and up to 300 participants online.
The ELG consortium has compiled a book that documents all aspects of the project and the initiative. The book consists of 30 chapters, the chapters were authored by the responsible work package partners and by the pilot project managers. The book had not been planned as part of the project work plan but was realized in addition to it. The ELG project coordinator is in the process of founding an ELG legal entity that will maintain the ELG infrastructure in the future. For this a transfer document is being prepared in which the details of the transition from ELG as a project to ELG as a registered association are regulated. The association is expected to be in place by the end of 2022.