Meetween

𝐂𝐨𝐦𝐞 𝐚𝐧𝐝 𝐦𝐞𝐞𝐭 #𝐌𝐞𝐞𝐭𝐰𝐞𝐞𝐧 𝐚𝐭 𝐋𝐑𝐄𝐂 𝟐𝟎𝟐𝟔! Two @FBK_research papers accepted. Palma de Mallorca, 11–16 May. #LREC2026

Two Meetween papers at #LREC2026 by @MaikeZufle, PhD researcher in instruction-following speech models and full-duplex dialogue at the Karlsruher Institut für Technologie (KIT). 📄 MuSaG 📄 Talk2Ref 👉 lrec2026.info #Meetween #NLP

4 papers accepted at #LREC -COLING 2026 in Palma de Mallorca 🎉 Sarcasm detection, reference prediction, gender bias in speech translation, and low-resource ASR. Meet the team, 11–16 May 2026 👉 lrec2026.info Follow us for paper deep-dives in the coming weeks. #NLP

Meetween is part of the organising committee of #IWSLT2026 — the premier conference on spoken language translation. The Shared Task Evaluation Period is open. Working on #speechtranslation, instruction following, or #modelcompression? Get involved now! 🔗 iwslt.org/2026

🎙️ Our paper on connecting Speech Foundation Models with LLMs is featured in the SpeechLMM Training Journal on Weights & Biases. Read it 👉 bit.ly/4svG7ll SpeechLMM 2.0 coming this summer. 👀 #Meetween #SpeechLMM #AI #NLP

🎙️ Two people. Two languages. One conversation!  No delays. No switching languages. No one is left out. This is what we are building. #SpeechAI #MultilingualAI #HorizonEurope

@sarapapi introduces MCIF, a new benchmark evaluating 23 SOTA models, revealing gaps in speech–video integration & long-form processing. 🎥 Watch the recording: youtube.com/watch?v=uQrX3i… #Meetween #MultimodalAI #LLMs #AIResearch

Meetween Butler enables real-time speech-to-speech translation while preserving voice and tone, plus automatic notes and smart summaries. Built on 𝐨𝐩𝐞𝐧-𝐬𝐨𝐮𝐫𝐜𝐞, 𝐩𝐫𝐢𝐯𝐚𝐜𝐲-𝐚𝐰𝐚𝐫𝐞 and 𝐫𝐞𝐬𝐩𝐨𝐧𝐬𝐢𝐛𝐥𝐞 𝐛𝐲 𝐝𝐞𝐬𝐢𝐠𝐧. #Meetween #ResponsibleAI

📢 Postdoc Opening in Multimodal AI I’m looking for a postdoc to join my lab at @itu1773 for our EU-funded @MeetweenEU project. You’d be working in a collaborative team with kind, smart, and hard-working people. Please feel free to share with anyone who might be interested.

#Mumospee Dataset v2 is out. 645,000+ hours of multilingual, multimodal meeting data, including 84,000+ new audiovisual hours, powering #SpeechLLMs, speech-to-speech translation and vision-based AI. Explore: bit.ly/4qIu8PC #Meetween #SpeechAI #MultilingualAI

Meetings where language isn’t a barrier and no insights are missed. 𝐁𝐮𝐭𝐥𝐞𝐫 𝐛𝐲 #𝐌𝐞𝐞𝐭𝐰𝐞𝐞𝐧 will 𝐭𝐫𝐚𝐧𝐬𝐥𝐚𝐭𝐞, 𝐭𝐚𝐤𝐞 𝐧𝐨𝐭𝐞𝐬, and 𝐬𝐮𝐦𝐦𝐚𝐫𝐢𝐬𝐞 𝐦𝐞𝐞𝐭𝐢𝐧𝐠𝐬 𝐢𝐧 𝐫𝐞𝐚𝐥 𝐭𝐢𝐦𝐞, all within your video calls. Clearer focus. Better conversations.

The Meetween consortium met in Rome for a two-day in-person session hosted by @Translation and @picampusschool. Partners aligned on SpeechLMM progress, roadmap, ethics and next steps toward open, trustworthy speech foundation models. #Meetween #EuropeanAI #OpenSourceAI

Two years in, Meetween is transforming how we meet. ✅ SpeechLMM v1 released ✅ 560k+ hrs of multilingual speech shared ✅ Global AI benchmarks + ethical design ✅ Built with real users The future of meetings is here, and it's inclusive. #Meetween #AIForGood

Online meetings connect people. They also add friction. Language. Time zones. Tools. Meetween is an EU-funded project exploring how AI can support natural and inclusive online collaboration. People first. Real conversations. Discover more: bit.ly/Meetween #Meetween #AI

IMTalker: Efficient Audio-driven Talking Face Generation with Implicit Motion Transfer TLDR: A motion transfer model, similar to LivePortrait, which replaces the typical warping paradigm with a cross attention model that implicitly learns motion. 📽️ Project Page: cbsjtu01.github.io/IMTalker/ 📜 Paper: arxiv.org/pdf/2511.22167 💻 Code: github.com/cbsjtu01/IMTal…

🔊 Meet FAMA: a 🇮🇹 Italian + 🇬🇧 English Speech Foundation Model, built in the MEETWEEN project under the #OpenScience framework! 👉 Check it out: arxiv.org/pdf/2505.22759

Weekly pick from the #MeetweenScientificWatch: Overtrained language models are harder to fine-tune — This study reveals how longer pretraining can harm downstream adaptability, introducing "catastrophic overtraining" as a real risk. 🧠📉 arxiv.org/abs/2503.19206

📢 The evaluation period of the Instruction Following task at @iwslt 2025 just started! 🖥️ Consider submitting your speech-to-text system! The outputs can be easily uploaded on the SPEECHM platform developed in the @MeetweenEU project! ➡️ iwslt2025.speechm.cloud.cyfronet.pl

Weekly pick from the #MeetweenScientificWatch: “The Ultra-Scale Playbook: Training LLMs on GPU Clusters” – A deep dive into scaling LLM training from 1 to 1000+ GPUs, demystifying parallelism techniques with practical examples. 🚀🖥️📖 huggingface.co/spaces/nanotro…

Treasure Hunt at the Meetween project meeting in @KITKarlsruhe, Karlsruhe last week! 🏴‍☠️🔍 Any guesses on what's inside the treasure? 👀💡 #Meetween #TreasureHunt #Karlsruhe