|
Newsletter #27 – May 2023
|
|
|
|
Dear reader,
META-FORUM 2023 is approaching, with less than two months left until the conference takes place on 27 June in Brussels. The programme will be made available soon on the website so make sure to register and stay tuned for more information!
We’re also happy to announce our new Digital Language Equality metric dashboard, featuring four new modes of data visualisation and language-comparison options.
Our featured ELG tool of the month is INCEpTION, an annotation tool for several semantic phenomena.
After the introduction of the nine selected FSTP projects in previous newsletter editions, the final project reports of six finished projects are now available on the ELE Open Call page.
In the section “From the SRIA”, we’re taking a look at the vision and recommendations for Speech Processing.
With best regards
Georg Rehm
|
|
|
Common European Language Data Space (LDS) Newsletter Launched
The European Language Data Space initiative that was started back in January just launched a monthly newsletter, providing information on the latest developments in secure, privacy-preserving language data sharing and use across Europe.
Subscribe to the newsletter for updates on LDS implementation, success stories, events, and more!
|
|
|
Language Technology and NLP in the news
|
|
|
- “European AI start-ups race to improve chatbots’ language skills” – Financial Times, 5 April 2023
-
“Deci’s NLP model clocks 100,000 queries per second in latest MLPerf results” – VentureBeat, 5 April 2023
-
“European privacy watchdog creates ChatGPT task force” – Reuters, 14 April 2023
-
“As AI agents like Auto-GPT speed up generative AI race, we all need to buckle up” – VentureBeat, 17 April 2023
-
“OpenAI’s CEO Says the Age of Giant AI Models Is Already Over” – Wired, 17 April 2023
-
“Elon Musk claims to be working on ‘TruthGPT’ — a ‘maximum truth-seeking AI’” – The Verge, 18 April 2023
-
“RedPajama replicates LLaMA dataset to build open source, state-of-the-art LLMs” – VentureBeat, 18 April 2023
-
“AI Regulation in Europe: Legal Challenges and Perspectives” – Lexology, 18 April 2023
-
“PlantGPT — the company building a generative model to speak the language of plants” – Sifted, 19 April 2023
-
“Stability AI announces new open-source large language model” – The Verge, 19 April 2023
-
“Inside the secret list of websites that make AI like ChatGPT sound smart” – The Washington Post, 19 April 2023
-
“Microsoft Research Propose LLMA: An LLM Accelerator To Losslessly Speed Up Large Language Model (LLM) Inference With References” – Marktechpost, 19 April 2023
-
“A dog, a horse, and a GPT large language model” – TechTalks, 20 April 2023
-
“Google’s big AI push will combine Brain and DeepMind into one team” – The Verge, 20 April 2023
-
“Large, creative AI models will transform lives and labour markets” – The Economist, 22 April 2023
-
“What is Auto-GPT and why does it matter?” – TechCrunch, 22 April 2023
|
|
Selected new tools and resources on the
European Language Grid
|
|
|
INCEpTION Text Annotation Platform - INCEpTION is an annotation tool for several semantic phenomena. It supports the user in assembling and compiling task-specific datasets.
|
|
|
Our next conference – META-FORUM 2023 – will take place on 27 June in Brussels, Belgium. We will present the final results of the European Language Equality project and discuss all kinds of topics touching upon language technologies, language resources, language-centric AI and especially digital language equality. We will talk about the future of the sector and also present the new ELE Book. You can register for free here.
The final project reports of six finished FSTP projects are now available on the ELE Open Call page, with three more coming up at the end of May.
|
|
|
A year after the launch of the interactive dashboard for the Digital Language Equality Metric on the ELG website, we are excited to announce a major update that brings new ways to visualise and compare language data. In addition to the previously available cross-language comparison, we have now introduced within-language comparison, allowing you to filter resources by datasets, resource subclasses, software, functions, and more within a single European language. We've also added heatmap and table options to display the number of resources for selected languages, making it easier to understand the distribution and availability of language resources across Europe. The new radial bar graph feature identifies the gaps and relevant factors necessary for further development of language technology for one or several selected European languages. Finally, track the progress of language technology over time with the option to display the evolution of the number of resources for selected languages. With these enhancements, we hope to provide an even more comprehensive and user-friendly overview of available tools, data, and resources, tracking each language’s digital readiness and contribution to the state of technology-enabled multilingualism and with that its progress towards the goal of Digital Language Equality.
|
|
|
Research Topic: Speech Processing
For the field of speech processing, several key recommendations should be considered. Prioritising the enhancement of speech resources and creating acoustic models for a wide variety of languages, including non-standard varieties and dialects, can provide comprehensive coverage for a diverse range of users. Developing natural-sounding synthetic voices is essential for enabling access to content in various spoken languages, while improving context modelling, especially for large volumes of text in translation tasks, can lead to more accurate translations. Addressing difficult audio conditions, such as multiple simultaneous speakers in noisy environments in different languages and tonality, is necessary for enhanced speech recognition and understanding. Supporting research combining speech, natural language understanding, and natural language processing with other modalities like image and vision can result in more comprehensive and integrated solutions. Lastly, tackling privacy and security threats related to speech synthesis, voice cloning, and speaker recognition is crucial for ensuring safe interactions with speech technologies while maintaining personal privacy and security.
You can read more about all SRIA recommendations here or take a look at the full document.
|
|
|
If you would like to voice your support for the ELE Programme and its goal and vision to achieve digital language equality in Europe by 2030, please consider filling out the endorsement form by clicking the button below and become a listed supporter on the ELE website:
|
|
|
- 3rd International Conference ‘Language in the Human-Machine Era’ (LITHME), 15-16 May, Groningen, Netherlands
-
META-FORUM 2023, 27 June, Brussels, Belgium
-
1st European Summer School on Artificial Intelligence (ESSAI) & 20th Advanced Course on Artificial Intelligence (ACAI), 24 July - 28 July, Ljubljana, Slovenia
-
34th European Summer School in Logic, Language and Information (ESSLLI), 31 July - 11 August, Ljubljana, Slovenia
If you have an event that you think the European language technology community should know about, get in touch with us to have it featured in this newsletter.
|
|
|
The next ELT newsletter will be sent out on 6 June 2023. Until then, follow our ELT social media accounts (as linked below) for the latest news!
|
|
|
|
|
