🚀 New Shared Task: Model Compression for Machine Translation at hashtag#WMT2026 (co-located with hashtag#EMNLP2026)! 📅 Test data out on June 18th, submissions by July 2nd! Can you shrink an LLM and keep translation quality high? 🧠🔧 👉 www2.statmt.org/wmt26/model-... #NLP #ML #LLM #ModelCompression

🎓 Today we attended the FBK PhD Day, and our amazing students presented their research lines: @apierg.bsky.social: gender-inclusive MT @linaconti.bsky.social: explainable speech translation @zhihangxie.bsky.social: long-form SpeechLLMs @uzumakidhairya.bsky.social: resource-efficient translation

🔍 Stiamo studiando come l'AI viene usata in Italia e per farlo abbiamo costruito un sondaggio! 👉 bit.ly/sondaggio_ai... (è anonimo, richiede ~10 minuti, e se partecipi o lo fai girare ci aiuti un sacco🙏) Ci interessa anche raggiungere persone che non si occupano e non sono esperte di AI!

🚀 New tech report out! Meet FAMA, our open-science speech foundation model family for both ASR and ST in 🇬🇧 English and 🇮🇹 Italian. The models are live and ready to try on @hf.co: 🔗 huggingface.co/collections/... 📄 Preprint: arxiv.org/abs/2505.22759 #ASR #ST #OpenScience #MultilingualAI

FAMA - a FBK-MT Collection

The First Large-Scale Open-Science Speech Foundation Model for English and Italian

huggingface.co

📢 Come and join our group! We offer a fully funded 3-year PhD position: 📔 Automatic translation with large multimodal models: iecs.unitn.it/education/ad... 📍Full details for application: iecs.unitn.it/education/ad... 📅 Deadline May 12, 2025 #NLProc #FBK

Interesting to see multimodal LLM built by combining modality encoders and LLM with adapters, as in the SFM+LLM paradigm, independently for each modality. This modularity may ease the creation of more MLMs from collaborations of single-modality experts. arxiv.org/abs/2501.04561

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

While we look forward to a sunny Geneva, why wait to join the conversation? We’ve created a starter pack for our #GITT2025 friends! 🕵️ Follow researchers working on gender bias in MT 💬 Stay up to date and dive into the discussion! All info at sites.google.com/tilburgunive...

very interesting to see more and more methods to close the length mismatch between speech and text sequences (aka length adapter -- see arxiv.org/abs/2402.12025) for SFM+LLM models! This one merging CTC and Q-former sounds very cool to me: arxiv.org/abs/2412.01145

AlignFormer: Modality Matching Can Achieve Better Zero-shot Instruction-Following Speech-LLM

Integrating speech into LLM (speech-LLM) has gaining increased attention recently. The mainstream solution is to connect a well-trained speech encoder and LLM with a neural adapter. However, the lengt...

arxiv.org

Next up: simultaneous speech translation! 🎯 Goal: to explore ways to translate speech into another language like simultaneous interpreting. 🔗 Link: iwslt.org/2025/simulta...

First up, a new task for 2025: *Instruction-following for speech processing!* Explore instruction-following for speech ⇨ Integrate speech foundation models with LLMs across tasks such as speech translation, recognition, summarization, and QA. 🔗: iwslt.org/2025/instruc...

Today's task: model compression!! 🆕 New at IWSLT! But no less exciting 🔥 🎯 Goal: Compress a large, general-purpose multimodal model, making speech translation more efficient ⚡️, deployable 📲, and sustainable ♻️, while preserving translation quality ⭐️ #AI #SpeechTech #ModelCompression #LLMcompression

I'm happy to share that our paper "Prepending or Cross-Attention for Speech-to-Text? An Empirical Comparison" has been accepted at @naaclmeeting.bsky.social 2025! #NAACL2025 @mgaido91.bsky.social 👏 📃 Preprint: arxiv.org/abs/2501.02370 ⏰ Code will be released soon #NLProc #Speech

Prepending or Cross-Attention for Speech-to-Text? An Empirical Comparison

Following the remarkable success of Large Language Models (LLMs) in NLP tasks, there is increasing interest in extending their capabilities to speech -- the most common form in communication. To integ...

arxiv.org

Hello world! 👋 We're coming out of hibernation to bring you this happy news: 1) We're organising the 3rd edition of GITT at #MTSummit! Working on #gender & #translation #technology? We'll see you there! 2) We're moving away from Twitter, so share the news and help us find old and new GITT friends!

a polar bear cub is laying in a pile of branches .

ALT: a polar bear cub is laying in a pile of branches .

media.tenor.com

Our #iwslt 2025 task on instruction-following speech models is out! Submission by April 15th. Check it out at: iwslt.org/2025/instruc...