Had such a great time presenting our tutorial on Interpretability Techniques for Speech Models at #Interspeech2025! 🔍 For anyone looking for an introduction to the topic, we've now uploaded all materials to the website: interpretingdl.github.io/speech-inter...

Roger Moore reminds Interspeech audience that speech is not audible text. Text is a technology.

And that's the whole team behind the tutorial on interpretability techniques for speech models at IS2025.

Wrap-up of the Interspeech tutorial on interpretability for speech models by @mdhk.net

Come join our tutorial in Rotterdam!

Feature attribution is always tricky. We show just how tricky it is for speech models.

The notifications are out and we have accepted a record-breaking number of papers! Of the 2332 papers sent for review, 1163 were accepted! Acceptance rate: 49.87% ⚠️ Camera-ready deadline: May 31! 🗓️ Registration opens soon 📨 Subscribe to our newsletter www.interspeech2025.org/newsletter

Looking forward!

Last week to apply for our PhD position!

@tilburg-university.bsky.social is on BSky. Let's get them to be more active and actually post something!

I’m again fucking begging anglo media to stop using „Caucasian” to mean „white” ESPECIALLY in any context involving Russia where oppression of actual Caucasians, as in people from ethnic groups inhabiting the Caucasus, is a real ongoing issue.

We listened to speeches and marched through the streets. There were free fries 🍟. Tilburgers were curious and friendly. Protests are not really a thing in the Netherlands 😁

🟥 Today we #strike! The Dutch govt - w/ support of #CDA & #CU - has chosen to break its promise of structural investments in universities and cut €1B+ from the budget. Full day of activities: Check out the teach-outs & join the main protest at 12.30, Willemsplein, Tilburg! @woinactie.bsky.social

My R package is in RopenSci and CRAN!🚨 🚨 🚨 🚨 *pangoling*: Access to Large Language Model Predictions This package is mostly for psycholinguists who just want to get the predictability or surprisal values of words without too much fuss (and without using […] [Original post on fediscience.org]

Modern LLMs "speak" hundreds of languages... but do they really? Multilinguality claims are often based on downstream tasks like QA & MT, while *formal* linguistic competence remains hard to gauge in lots of languages Meet MultiBLiMP! (joint work w/ @jumelet.bsky.social & @weissweiler.bsky.social)

I'm now searching for new PhD students who are interested in studying how people's brains understand sequences of images like in comics. So, a mix of comics, cogneuro (EEG), and psycholinguistics (especially syntax/semantics). Please circulate to all who might be interested! tiu.nu/22739

Speech people! The latest #ISCApad is out: services.isca-speech.org/iscapad/isca... With great news: the ISCA medalist 2025 is @rogerkmoore.bsky.social 😍 Super happy for him! Roger was my first post-doc advisor. I'm very much looking forward to handing him the Medal at @interspeech.bsky.social 🥳

نوروز مبارک Happy Nowruz

📢 #SLaTE2025 – Call for Papers! 🎤📚 🚀 10th Workshop on Speech & Language Technology in Education 📅 Aug 22-24, 2025 | 📍 Radboud Univ., Nijmegen 📝 Special Session: Speak & Improve Challenge 2025 📄 Submit full/short papers 🔗 Info: sites.google.com/view/slate-2...

u24.gov.ua Best what one can do as individual is it set a monthly target and donate it until the victory automatically. For example as standing order from your bank using SEPA: koloua.com/en/donate#sepa

I just donated to u24.gov.ua

Still thinking about this young mathematician. Her name was Yulia Zdanovska. She could have done so much for her country and for our shared world. We all lost her. Killed by Putin’s Russia in Kharkiv almost exactly three years ago. Memorial from the ACM: cacm.acm.org/news/in-memo...

I just donated to u24.gov.ua

Radek Sikorski addresses the UN on the war in Ukraine. www.reddit.com/r/europe/s/t...

The best way to confront a bully about the misinformation he is spreading is to confront him in front of everyone. Macron just did that.

I want to tell you a story. In 2015, I was in the small Lithuanian city of Panevežys, standing with the locals, watching a military parade. But this was no ordinary parade, because marching alongside the Lithuanian troops were our American allies.🧵1/17

We are looking for reviewers for Interspeech 2025. Consider signing up at: isca-speech.org/Reviewing

🔴 Paper submission platform for #Interspeech2025 is now open! www.interspeech2025.org/paper-submis...

Note that for the first opening, relevant fields of expertise we're looking for include Language & Speech Processing and Deep Learning for Perception.

Our department (Cognitive Science and AI at Tilburg University) has two openings for new Assistant Professors: one in Data Science and AI, and one in Cybersecurity. Consider joining us. www.academictransfer.com/en/347847/as... www.academictransfer.com/en/347846/as...

Special Session: Biosignal-Enabled Spoken Communication @ #Interspeech2025 💡 From EEG to EMG, biosignals are revolutionizing speech communication! Explore cutting-edge applications in speech restoration, prostheses, & more. 🔗 Details: www.uni-bremen.de/csl/interspe...

💢Special Session: Challenges in Speech Data Collection, Curation, and Annotation at #Interspeech2025 📊 Speech Data is the backbone of innovation. Share lessons & explore new workflows for better data collection, curation, & annotation. 🔗 Learn more: sites.google.com/view/speech-...

Phase Transition xkcd.com/3025

Source Tracing: The Origins of Synthetic or Manipulated Speech at #Interspeech2025 🔍 Is a speech sample real or synthetic? Join the Special Session to advance tools for detecting manipulated speech. 🎯 Topics: Forensics, cybersecurity, ethics, & xAI 🔗 deepfake-total.com/sourcetracing

⚠️Special Session: Responsible Speech Foundation Models 🌐 Explore ethical, fair, & secure use of Speech Foundation Models at #Interspeech2025! Dive into de-biasing, robustness, & multimodal approaches for responsible innovation. 🔗 Details: sites.google.com/view/respons...

🎭 Speech Emotion Recognition in Naturalistic Conditions Challenge @ #Interspeech2025 Advance SER from spontaneous speech with MSP-Podcast corpus: 324+ hrs conversational data Tasks 1️⃣ Categorical Emotion Recognition 2️⃣ Emotional Attributes Prediction 🚀 lab-msp.com/MSP-Podcast_...

Most students don't really like writing so they outsource it to chatbots. Most teachers don't really like grading student essays so they outsource it to chatbots. Win-win?

Looking for reviewers before Christmas

📜 The author kit is now available on interspeech2025.org!

🌟 URGENT Challenge @ #Interspeech2025 🌟 Join the Universal, Robust, & Generalizable Speech EnhancemeNT (URGENT) challenge! Explore noisy corpora, tackle diverse speech degradations, and test scalability across 2 tracks (~2.5k/60k hrs). 🚀 Learn more: urgent-challenge.github.io/urgent2025/

Justified smugness sounds like fun. But I'm not so sure what ate what in this case. It's not so long ago that NLP was all about POS tagging, parsing, word sense disambiguation and other miscellaneous pursuits which seem rather unfathomable to us today.

Our piece is finally out in the Imminent blog! 🎉It presents preliminary findings of our recent study evaluating the usefulness of word-level quality estimation in real-world post-editing settings (paper forthcoming)! 🧵1/ imminent.translated.com/can-word-lev...

Very true. I kind of regret coming up with the name BlackboxNLP for the workshop. Though GlassboxNLP wouldn't have been nearly as catchy.

Neural networks are not black boxes. They're the opposite of black boxes: we have extensive access to their internals. I think people have accepted this framing so innately that they've forgotten it's not true and it even warps how they do experiments.

Madrid tripled the length of its metro system in just 12 years, faster and cheaper than almost any other city in the world. Left: in 1994 Right: after 2007 Cost per mile: £64–91 million (around 7x cheaper than London's Jubilee & Northern line expansions & 14x cheaper than Toronto's Ontario line)

🐭🔊 MAD-UV Challenge at #Interspeech2025 Innovate at the intersection of speech tech & biomedical research! Develop models to classify ultrasonic vocalizations from ASD & wild-type mice. Advance autism research using cutting-edge bioacoustic analysis. www.mad-uv.org

My PhD thesis is now available online! It is called "Finding Structure in Language Models" and presents the work I have done in the past 4 years on the intersection of NLP, interpretability, and linguistics. I'll defend it on December 10, streamed here: jumelet.ai/phd arxiv.org/abs/2411.16433 🧵

Nice to see this article by @mbaldwin.bsky.social surface again. Lots of academics aren't aware how recent and dynamic the peer-review system is. Entire thing worth a read, but this quote always floors me (page 8 of PDF ethos.lps.library.cmu.edu/article/19/g... ):

Another question to the speech community here: Which paper is considered to be the first paper on speech recognition that used DNNs? This one: Acoustic Modeling using Deep Belief Networks Abdel-rahman Mohamed, George E. Dahl, and Geoffrey Hinton? IEEE TASLP or another...?