|
Newsletter #21 – November 2022
|
|
|
|
Dear reader,
the first version of our Strategic Research and Innovation Agenda (SRIA) towards digital language equality in Europe by 2030 is steadily getting closer to being completed. Starting with the next edition, we will launch a new section in the newsletter in which we will have a closer look at our recommendations to reach digital language equality by 2030.
A week from now, the STOA workshop on digital language equality is happening in Brussels (as well as via an online stream), so get ready to register if you want to attend and join the discussion. This newsletter is also the last one before the 29 November submission deadline for our Open Call for SRIA Contribution Projects.
Featured in our tools and resources section is a GPT-2-powered interactive theatre play scene script generator.
In the ELG section, we’re taking a breather to have a look at the current numbers, as well as a reminder to register, use and contribute to the grid.
With best regards
Georg Rehm
|
|
|
Language Technology and NLP in the news
|
|
|
- “How to choose the right NLP solution” – VentureBeat, 1 October 2022
-
“The Problem With Mental Health Bots” – Wired, 1 October 2022
- “How will OpenAI’s Whisper model impact AI applications?” – VentureBeat, 3 October 2022
- “Amazon, Apple, Microsoft, Meta and Google to improve speech recognition for people with disabilities” – Engadget, 3 October 2022
- “Get ready for the next generation of AI” – MIT Technology Review, 3 October 2022
- “The nuances of voice AI ethics and what businesses need to do” – VentureBeat, 8 October 2022
- “How AI and machine learning are changing the phishing game” – VentureBeat, 10 October 2022
- “Why AI Will Never Fully Capture Human Language” – Sapiens, 12 October 2022
- “AI Voice Assistants Could Now Become Amazing Language Teachers” – Towards AI, 13 October 2022
- “AI language models could be offensive or prejudiced toward individuals with disabilities” – News Medical, 13 October 2022
- “AI is changing scientists’ understanding of language learning – and raising questions about an innate grammar” – The Conversation, 19 October 2022
- “Meta AI announces first AI-powered speech translation system for an unwritten language” – VentureBeat, 19 October 2022
- “AI technology is not dark magic, it’s just misunderstood” – Open Access Government, 20 October 2022
- “T2CI GAN: A deep learning model that generates compressed images from text” – TechXplore, 26 October 2022
|
|
|
In lieu of our usual news, we’re taking the time to look at what 2 and a half years of contributions to the European Language Grid have brought us: 6974 corpora, 3667 tools & services, 2714 conceptual resources, 482 models & grammars as well as 1775 listed organisations.
If you haven’t registered yet and want to contribute and make use of our collective resources, follow the link to the registration page. Feel like you need a hand getting started? Be sure to check out our in-depth video tutorial on how to use the ELG.
|
|
Selected new tools and resources on the
European Language Grid
|
|
|
|
THEaiTRobot 2.0 - The THEaiTRobot 2.0 tool allows the user to interactively generate scripts for individual theatre play scenes. The previous version of the tool was based on the GPT-2 XL generative language model, using the model without any fine-tuning, as we found that with a prompt formatted as a part of a theatre play script, the model usually generates a continuation that retains the format. The current version also uses vanilla GPT-2 by default, but can switch to a GPT-2 medium model fine-tuned on theatre play scripts (as well as film and TV series scripts). Apart from the basic "flat" generation using a theatrical starting prompt and the script model, the tool also features a second, hierarchical variant, where in the first step, a play synopsis is generated from its title using a synopsis model (GPT-2 medium fine-tuned on synopses of theatre plays, as well as film, TV series and book synopses). The synopsis is then used as input for the second stage, which uses the script model. The choice of models to use is done by setting the MODEL variable in start_server.sh and start_syn_server.sh THEaiTRobot 2.0 was used to generate the second THEaiTRE play, "Permeation/Prostoupení".
|
|
It’s almost time for our STOA workshop on digital language equality! On November 8, the workshop will explore the research and development environment of language technologies in the context of multilingualism in the European Union. Find the link to register and have a look at the full programme here. If you feel like diving into the matter right away, check out this STOA podcast and article titled “What if everyone spoke the same language?”.
Moreover, the pre-final recommendations of our Strategic Research and Innovation Agenda will be presented and discussed in the STOA workshop. Starting with the next edition of the ELT newsletter, there’s going to be a monthly look into the SRIA, each dedicated to a different chapter of the roadmap, starting with the scientific goal of reaching Deep Natural Language Understanding by 2030.
The deadline for our Open Call for SRIA Contribution Projects is also approaching, with only four more weeks until November 29, so be sure to register and submit your proposals if you want to participate.
|
|
|
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 December 2022. Until then, follow our ELT social media accounts (as linked below) for the latest news!
|
|
|
|
|
