Dear reader,

today we start into an exciting February: next week, the French Ministry of Culture hosts its three-day online forum on “Innovation, Technologies and Plurilingualism”, which includes a workshop on the European Language Grid. Unfortunately, a small typo had found its way into the description of the registration process featured in our last ELT newsletter – check out the ELG section below for further information and the correct way to register for the event.

Last week, a keen interest regarding multilingualism was registered by ECSPM President Bessie Dendrinos when she gave her lecture on EU policies and practices at a seminar series of the European Civic University (CIVIS). Students and postgrads from four European universities participated and learned about the ELE and ELG projects and their link to EU language policies and funding strategies, such as the right of all EU citizens to use the official languages of the EU Member States. You can learn more about the lecture in the ELE section below.

Newsletters are commonly a one-directional form of communication. However, when it comes to the European Language Grid and its user experience, we would like to initiate a dialogue and hear from you: how do you use the ELG?  Are there functionalities that you enjoy? Or features that are missing? We would like to collect your feedback and suggestions via email and will take it into account  in the further development of the platform.

We hope you enjoy our newsletter and have a good week!

With best regards

Georg Rehm

• “Facebook Open-Sources Two Billion Parameter Multilingual Speech Recognition Model XLS-R” – InfoQ, 18 January 2022
• “Google Research Brings ‘Massively Multilingual’ Machine Translation to 200+ Languages” – Slator, 18 January 2022
• “The fast and the curious: New Dutch VC unveils €50 million fund for AI startups in Northern Europe” – Tech.ui, 19 January 2022
• “From Age-Gating to Chapter Indexing – How YouTube Uses AI” – Dataconomy, 20 January 2022
• “For Dataset Creation, Humans Should Evaluate Machine Text — New Research” – Slator, 20 January 2022
• “Adopting neural language models for the telecom domain” – Ericsson, 20 January 2022
• “AI models are becoming better at answering questions, but they’re not perfect” – VentureBeat, 21 January 2022
• “Meta Unveils New AI Supercomputer” – The Wall Street Journal, 24 January 2022
• “Heidrick & Struggles data shows diversity in AI needs work” – Sustainability Magazine, 26 January 2022
• “OpenAI rolls out new text-generating models that it claims are less toxic” – VentureBeat, 27 January 2022
• “EU Commission wants to set world's first digital standards” – EUobserver, 27 January 2022

• Do Greek words used in English make politicians sound smart? The “former Greek Language Minister” (enacted by Yiannis Vassilakis) has a response.
• Language technology and “well-being” are not always used in the same sentence – but when they are, they make up for a good quote.
• Did you know that words for “rough” sound rough in many languages around the globe?

Less than a week is left until the start of the online forum on “Innovation, Technologies and plurilingualism” hosted by the French Ministry of Culture within the framework of the French Presidency of the Council of the European Union. The three-day event kicks off on Monday, 7 February 2022, and includes a workshop on the European Language Grid on Tuesday, 8 February 2022. The full programme can be downloaded here.

Some of our readers discovered a typo that had found its way into our last newsletter regarding the registration process: after creating an account on the platform of the French EU Council Presidency and activating it via the email that is sent out afterwards, you need to register in the delegation FORINOVPLURI.PARTICIDGLFLF – with a period (.) in the centre instead of a colon (:). The access code for the delegation is ParticipaDGLFLF2022! with the exclamation mark being part of the code. We are sorry about the confusion and look forward to seeing you all online next week.

Further event news concern the upcoming national ELG Workshop hosted by the Department of Computational Linguistics at the Bulgarian Academy of Sciences in Sofia, Bulgaria: The First Bulgarian dissemination event in the context of the European Language Grid (ELG) takes place online on 11 February at 10am and is aimed at a specific audience: On the website of the DCL, you can find five different questions regarding your work with language technology – if your answer to at least one of them is “yes”, make sure to attend!

Finally, we take a look back at last year’s META-FORUM 2021: We invited the ten pilot projects that were funded in the first ELG Open Call to produce short videos for the online conference. The videos were shown during the pilot project session and at the virtual expo. Now, they are finally available on YouTube! Have a look at our overview on the ELG website and revisit these great projects, such as the Basque-speaking smart speaker or the virtual personal assistant YouTwindi.

GATE Hate – A service that tags abusive utterances in any text. It includes a feature, "type", indicating the type of abuse if any, such as sexist, racist etc., and a "target" feature that indicates if the abuse was aimed at the addressee or some other party. The service will also tag a range of politically relevant topics, as well as entities such as persons, locations and organizations and Twitter entities such as hashtags and user mentions. The running tool is available in ELG, therefore, everybody can directly try it out with an example sentence on any English language text. The tool was added by Ian Roberts from the University of Sheffield and published in our catalogue on the 20 January 2022.

“Multilingualism: EU policies and practices” is the title of a lecture delivered on 25 January 2022 by Prof. Bessie Dendrinos, ECSPM President, who referred to the European Parliament’s resolution on language equality in the digital age. She briefly described the European Language Equality (ELE) project and its purpose, but also explained the link between this project and the European Language Grid (ELG) on its way to being the primary platform for the Language Technology (LT) community in Europe. She also showed the connections between ELE and ELG with the European Language Resource Coordination (ELRC), META-NET and CLARIN, all of which may be of particular interest to research institutions.

The lecture was delivered during the second session of an online seminar series entitled “Multilingualism: The Language of CIVIS”, organised by the Multilingualism Network functioning in the framework of CIVIS – a European Civic University formed by the alliance of 10 leading research higher education institutions across Europe. It was attended by seniors and postgraduates in language study programmes, doctoral and postdoctoral students of the universities of Athens (NKUA), Madrid (UAM), Stockholm and Glasgow, who showed a keen interest in the forms of governance, language policies, resolutions and recommendations developed at an increasing pace since the early 90s in order to secure EU’s multilingualism, which is one of its founding principles. It also presented the strategies and tools which have been developed to secure the fundamental right of all EU citizens to use the official languages of the Member States, but also to protect and promote regional, minority and endangered European languages to safeguard the diversity of the Union’s linguistic and cultural heritage (EU Parliament resolution 2013).

• 7-9 February: Innovation, Technologies and plurilingualism (online forum)
• 11 February: First Bulgarian dissemination event in the context of the European Language Grid (ELG) (online)

HENSOLDT Analytics

HENSOLDT Analytics GmbH (previously SAIL, “Speech, Artificial Intelligence and Language Laboratories”) is a leading global provider of automatic speech recognition (ASR) and natural language processing (NLP) technologies. SAIL was founded in 1999 in Vienna, Austria and acquired by HENSOLDT in January 2021, leading to a change of name to HENSOLDT Analytics. It is now based in Vienna, Austria as well as in Taufkirchen, Germany.

HENSOLDT Analytics develops technologies and solutions to mine traditional as well as social media contents in multiple languages from a wide portfolio of sources. This includes the indexing of audio, video, images and text by employing automatic speech recognition, various NLP technologies (such as named entity recognition, topic detection or sentiment analysis) as well as a set of visual processing technologies (face detection and recognition, object detection and tracking).

All technologies, models and analytical capabilities aim to combine traditional and social media and have been created with a cross-media, cross-platform and multilingual mindset. Harvesting and enrichment technologies are complemented by backend-technologies based on state-of-the-art big-data technologies to store media and make them accessible for rapid visualisation and advanced analysis purposes.

In our globalised economy, HENSOLDT Analytics has long been involved in the processing of sources from around the World providing contents in numerous languages. Starting out as a European SME and competing with global players (predominantly from the U.S.), HENSOLDT has always subscribed to the importance of providing multilingual technologies and models, able to support decision makers in processing a wide variety of contents. Especially in security-related matters, being able to tap into communication in many different languages is of the essence – be it to increase situational awareness during campaign monitoring, disaster management or supporting the fight against mis- and dis-information. In addition to large official languages, HENSOLDT emphasises the importance of being able to process data in lesser spoken languages as well as dialects.

HENSOLDT Analytics CSO (Chief Scientific Officer) Gerhard Backfried states: “We are delighted and honoured to take part in the ELG and ELE projects as we truly believe that language resides at the core of AI in many domains. Multilingual technologies, models and services are of the essence in many domains and target markets including the security and defence industry. Furthermore, these projects and their results help to strengthen European technologies and players in a competitive and worldwide market!”

The next ELT newsletter will be sent out on 15 February 2022. Until then, follow our ELT social media accounts (as linked below) for the latest news!