European Language Grid Platform

Towards the Primary Platform for Language Technologies in Europe

Our ambition is for the European Language Grid to become the primary platform for the whole European Language Technology landscape.

ELG aims at listing all relevant stakeholders, from technology development to research centres, from small and medium-sized companies to large enterprises. Stakeholders will be able to register themselves with corresponding entries in the ELG , which will act like the Yellow Pages of European Language Technology.

Through the ELG , companies, academic organisations and individual researchers can gain visibility and connect to each other – they will also be able to connect with the services, data sets and resources they decide to make available through the platform.

The resulting online catalogue will be browsable, searchable and explorable. Users will be able to filter and search for domains, sectors, regions, countries, languages, service types, data sets and more.


Technical Content of the European Language Grid

The platform will be populated with four abstract groups of technical content: functional Language Technology services, functional remote Language Technology services, non-functional tools and software packages (including data sets) as well as information about the wider Language Technology landscape including stakeholders.

Services and tools

Users of the ELG will be enabled to register, describe, upload, search and deploy as well as integrate containerised services, from simple tokenisation or part-of-speech tagging to complex processing workflows.

Initially be populated with

GATE , GATECloud , UDPipe, TILDE’s and University of Edinburgh’s MT Services, SAIL LABS ASR , KWS , sentiment, age and gender detection tools, etc.

Information about services, tools, resources

Non-functional catalogue entries, for example, on a Language Technology provider company or research centre.

Initially be populated with

Data sets, resources, models, tools from META -SHARE, ELRC -SHARE, and ELRA catalogue, ELRA ’s services for e-licensing and e-Commerce

Remote services and tools

Users of the ELG will be enabled to register, describe, search and integrate functional remote APIs.

Initially be populated with

Research and commercial services, among others, GATECloud with 65+ services, UDPipe, TILDE’s and University of Edinburgh’s MT Services, SAIL LABS and EXPERT SYSTEM services

Language resources, services, tools, software code

Users will be enabled to upload, describe, search, download data sets, source code packages, trained models, embeddings etc.

Initially be populated with

Data sets, resources, models, tools from META -SHARE, ELRC -SHARE, and ELRA catalogue, ELRA ’s services for e-licensing and e-Commerce


Platform Architecture

The European Language Grid will be the first large-scale Language Technology platform applying containerisation, using Docker in combination with Kubernetes and OpenStack. For LT , we consider containers to be a solution for the notorious issue of providing interoperability between different services. Operating system, programming languages, frameworks and library dependencies can all be included in a single image, ensuring that the service will operate as foreseen by its developer.

The ELG cloud infrastructure will be able to receive, manage and run many images in parallel. This is especially relevant for those organisations who cannot or do not want to host tools or services themselves (because, for example, they don’t have any infrastructure).


Timeline

The following timeline presents the most important phases and milestones of the European Language Grid project.

ELG timeline