Hi again! Please find me in these better places: .🦋 @/mdhk.net .🐘 @/mdhk@scholar.social .💼 👋.

The schedule features presentations and interactive sessions with a great team of co-organizers: @pouwcharlotte, @linguisticshen, Martijn Bentum, Tom Lentz, @hmohebbi75, Willem Zuidema, @gchrupala (and me!) We look forward to seeing you there 😃.

The @ISCAInterspeech early registration deadline is coming up in a few days!. Want to learn how to analyze the inner workings of speech processing models? 🔍 Check out the programme for our tutorial:.& sign up through the conference registration form!

Join the protest: Sign the Dutch petition (from NL): .Sign the international petition to show support from abroad: Find more materials and local action groups here:

In a bizarre undemocratic turn of events, the massive national protest against our government's plans for higher education was cancelled last week. We'll be back stronger next Monday in The Hague! 🟥

RT @mariannedhk: ✨ Do current neural speech models show human-like linguistic biases in speech perception?. We took inspiration from classi….

Or this from within The Netherlands: .and come to Utrecht on Nov 14th!

The Dutch government has planned severely destructive budget cuts to higher education and research funding. Please protest with us! 🟥. Consider signing this petition to show support from abroad:

RT @EvaPortelance: New paper out 🎉 !! @linguistMasoud and I wrote a perspective on how modern LMs and AI models can be useful for studying….

RT @RTomMcCoy: 🤖🧠NOW OUT IN PNAS🧠🤖. Language models show many surprising behaviors. E.g., they can count 30 items more easily than 29. In E….

RT @AdityaYedetore: NEW PAPER! We (@najoungkim and I) find that training on mapping from form to meaning leads to improved hierarchical ge….

RT @aakashagr92: 🚨 Publication Alert! .#PLOSCompBio: Cracking the neural code for word recognition in convolutional neural networks https:….

I will be presenting this work tomorrow (Thursday) at #INTERSPEECH2024, 10.00-10.40 in the Acesso room! . Looking forward to discuss how we can learn from human speech science to interpret end-to-end neural speech models 💡. The paper is here:

RT @oskarvanderwal: Working on #bias & #discrimination in #NLP? Passionate about integrating insights from other disciplines? Want to discu….

RT @spiantado: So much new and exciting work -- from LLMs to VSAs -- reshapes the science of concepts. See our new @TrendsCognSci review on….

RT @asifa_majid: The Open Encyclopedia of Cognitive Science is now live! OECS an online, open resource about the na….

This paper is the result of Bram's first research internship in the @MBCS_UvA programme, from which he is about to graduate with some very interesting follow-up work — it has been a joy to supervise this project and I am looking forward to the next steps!.

It also holds for sentences of different complexities! However, we observe an interesting difference in the relation of syntactic complexity to probed dependency accuracy 📉 vs. brain alignment 📈. ➡️ Come to our talk or read the paper to find out more!

We find this pattern across a range of mono- and multilingual language models processing Dutch sentences, measuring similarity to fMRI activity recorded in Dutch participants reading those same sentences (obtained from a large open dataset .

It turns out the accuracy of dependency structures decoded from LM hidden layers (measured by Labelled Attachment Score) strongly correlates with similarity to brain activity in sentence reading! 🧠 .This correlation disappears in a control condition with scrambled inputs.