news

Sep 08, 2026 📄 2 new papers (related to speech models) accepted at EMNLP workshops!
  • Abdu’s latest PhD work titled “Align, Integrate, and Fire: Efficient Token-Level Alignment for Zero-Shot SpeechLLMs” will be presented at WMT
  • Nicolas Bourrel’s MSc thesis titled “A Probe Shift Is Not a Fairness Fix: The Limits of Representation Steering in Speech Models” will be presented at IMPACT-SPEECH. Pre-prints soon!
Sep 04, 2026 ⚖️ JUST-IN kick-off meeting on October 8th. Stay tuned for more info!
Sep 03, 2026 🏳️‍🌈📊 New pre-print: WinoQueer-NL brings the first Dutch-language benchmark (based on a community survey) for measuring anti-queer bias in AI language models (work led by master student Jiska Beuk). Pre-print here, data here
Jun 10, 2026 📄 New paper accepted at INTERSPEECH 2026, titled “Cross-Modal Robustness Transfer (CMRT): Training Robust Speech Translation Models Using Adversarial Text”. Paper was led by Abdu. Pre-print here.
Feb 10, 2026 🎉 New grant awarded: We receive funding under HORIZON-CL2-2025-01-DEMOCRACY-11 for JUST-IN project (Justice-in-Context: Social practices, local realities and the everyday work of judicial independence in Europe) to start in October 2025. Will be looking for a postdoc in Law&AI, get in touch if interested!