Francesca Padovani

Thrilled to be heading to Suzhou with a big team of GroNLP'ers 🐮

Interested in Interpretable, Cognitively inspired, Low-resource LMs? Don't miss our posters & talks #EMNLP2025!

...ii) a direct quality reward from a teacher model, and iii) a reward based on the log probabilities of a teacher model (and its dialogue continuations). While these rewards did not improve our models performance, two different DPO approaches did!

Dialogue Is Not Enough to Make a Communicative BabyLM (But Neither Is Developmentally Inspired Reinforcement Learning) Francesca Padovani1∗ Bastian Bunzeck2∗ Manar Ali2 Omar Momen2 Arianna Bisazza1 Hendrik Buschmeier2 Sina Zarrieß2 1Center for Language and Cognition (CLCG), University of Groningen 2CRC 1646 – Linguistic Creativity in Communication, Bielefeld University f.padovani@rug.nl bastian.bunzeck@uni-bielefeld.de

As part of this year's BabyLM challenge, we (researchers from @gronlp.bsky.social and @clausebielefeld.bsky.social diverged from established pretraining paradigm by training only on dialogue data from CHILDES.

Don’t hesitate to reach out with any questions or doubts :)
Every contribution is more than welcome! 🙏😍

These are living datasets that researchers around the world can enrich as new resources become available
On the website of the project babylm.github.io/babybabellm/
you can find information about the dataset creation pipeline, training, and evaluation.

This was made possible thanks to the dedication of each language-specific expert and the coordination of senior researchers, @jumelet.bsky.social and @lchoshen.bsky.social among others.

𝐃𝐨 𝐲𝐨𝐮 𝐫𝐞𝐚𝐥𝐥𝐲 𝐰𝐚𝐧𝐭 𝐭𝐨 𝐬𝐞𝐞 𝐰𝐡𝐚𝐭 𝐦𝐮𝐥𝐭𝐢𝐥𝐢𝐧𝐠𝐮𝐚𝐥 𝐞𝐟𝐟𝐨𝐫𝐭 𝐥𝐨𝐨𝐤𝐬 𝐥𝐢𝐤𝐞? 🇨🇳🇮🇩🇸🇪

Here’s the proof! 𝐁𝐚𝐛𝐲𝐁𝐚𝐛𝐞𝐥𝐋𝐌 is the first Multilingual Benchmark of Developmentally Plausible Training Data available for 45 languages to the NLP community 🎉

arxiv.org/abs/2510.10159

This was made possible thanks to the dedication of each language-specific expert and the coordination of senior researchers, in particular @jumelet.bsky.social ! 🤩

Computational Psycholinguistics Meeting 2025 | Universiteit Utrecht Universiteit Utrecht

Computational Psycholinguistics Meeting 2025

cpl2025.sites.uu.nl

When: December 18–19, 2025

Where: Utrecht, the Netherlands

Abstract submission deadline: June 15, 2025

Organizers: Jakub Dotlačil, Lena Jäger, Bruno Nicenboim, Ece Takmaz

Jane Goodall (1934-2025) 🤍

My very first book review is out now 📚
Muchas gracias to @stefanhartmann.bsky.social for inviting me, looking forward to our next project(s) 😇

𝐓𝐡𝐞 𝐈𝐍𝐂𝐑𝐄𝐂 𝐩𝐫𝐨𝐣𝐞𝐜𝐭 (𝐄𝐑𝐂 𝐂𝐨𝐆, 𝐏𝐈 𝐀𝐧𝐚 𝐆𝐮𝐞𝐫𝐛𝐞𝐫𝐨𝐟) 𝐢𝐬 𝐥𝐨𝐨𝐤𝐢𝐧𝐠 𝐟𝐨𝐫 𝐚 𝐧𝐞𝐰 𝐭𝐞𝐚𝐦 𝐦𝐞𝐦𝐛𝐞𝐫! ⁣🎉
⁣
We have an opening for a 𝐏𝐨𝐬𝐭𝐝𝐨𝐜 𝐩𝐨𝐬𝐢𝐭𝐢𝐨𝐧 in 𝐂𝐫𝐞𝐚𝐭𝐢𝐯𝐢𝐭𝐲, 𝐓𝐫𝐚𝐧𝐬𝐥𝐚𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐓𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲 for a starting duration of 12 months extendible to 30 months.⁣

For further details -> www.rug.nl/about-ug/wor...

𝗧𝗵𝗲 𝗘𝘂𝗿𝗼𝗽𝗲𝗮𝗻 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗲𝗿𝘀’ 𝗡𝗶𝗴𝗵𝘁 𝘄𝗮𝘀 𝗮𝗻𝘁𝗶𝗰𝗶𝗽𝗮𝘁𝗲𝗱 𝗯𝘆 𝗮𝗻 𝗲𝗾𝘂𝗮𝗹𝗹𝘆 𝗲𝘅𝗰𝗶𝘁𝗶𝗻𝗴 𝗮𝗳𝘁𝗲𝗿𝗻𝗼𝗼𝗻! 😍
Here are some photos of our GroNLP station at the Forum, where we welcomed high school classes and potential future researchers. 🧑‍🔬🤖
Quizzes, puzzles, demos, card games, and of course, lots of chocolate! 🍫

Thanks to the entire CIRM team and the @ilcb.bsky.social for organizing this truly special week !

I can only 𝗵𝗶𝗴𝗵𝗹𝘆 𝗿𝗲𝗰𝗼𝗺𝗺𝗲𝗻𝗱 𝗽𝗮𝗿𝘁𝗶𝗰𝗶𝗽𝗮𝘁𝗶𝗻𝗴 in this training week in its future editions! 🌳

A special thanks to my amazing colleagues from Amsterdam, @mdhk.net and @cpouw.bsky.social , and to Felicia Stich from Göttingen :))

I had the chance to meet fellow researchers full of brilliant ideas and a genuine willingness to share, attend lectures, and listen to invited speakers who were not only knowledgeable but eager to engage in active intellectual exchange.

Set in a breathtaking location amid the Mediterranean scrub of Parc des Calanques, the program was intense and incredibly enriching.

The 𝗜𝗟𝗖𝗕 𝗦𝘂𝗺𝗺𝗲𝗿 𝗦𝗰𝗵𝗼𝗼𝗹 in Marseille went beyond all my expectations! 💯

A week has already flown by since I had one of the most formative experiences of my PhD so far. 👩‍🎨

We recharged our batteries with a wonderful weekend together in the province of Drenthe! 🌳🍃
As per tradition, we kick off the academic year with plenty of sharing, smiles, and sports 💘

We wish everyone a wonderful start of the academic year <3 may it be inspiring and exciting 🎓

Rowena Garcia is presenting research at #ISB15 showing that #crosslinguistic #priming occurs across #typologically unrelated languages & without overlap of constituent order. The degree of overlap modulated the magnitude of the effect.
#Tagalog #Indonesian #English

@isbilingualism.bsky.social

#ACL2025 has just finished and it was great to meet with old friends and connect with new colleagus. In case you missed our papers/presentations, this was us 🐮👇
@facultyofartsug.bsky.social

Children & Digital Technology | The Controversies In this guide we discuss and make practical recommendations for parents on children's use of digital technology.

Fabulous blog about screentime from Gonzales, Golinkoff & Hirsh-Pasek here 1/ childandfamilyblog.com/children-and...

Constructing language: a framework for explaining acquisition Explaining how children build a language system is a central goal of research in language acquisition, with broad implications for language evolution, adult language processing, and artificial intelli...

Children are incredible language learning machines. But how do they do it? Our latest paper, just published in TICS, synthesizes decades of evidence to propose four components that must be built into any theory of how children learn language. 1/
www.cell.com/trends/cogni... @mpi-nl.bsky.social

Excited to welcome a new member to the family of language-specific BLiMPs 🎊 this time putting Turkish in the spotlight!
Thanks to the meticulous, linguistically grounded work of the brilliant @ezgibasar.bsky.social. We hope will help push forward research on typologically diverse languages!

CDL may still hold unique value when used in models that better reflect the dynamics of real-world language learning 🤗 🤲

Let's keep investigating!

Trained on static text without interaction or feedback, they miss the communicative and social context central to how children learn. Rather than dismissing CDL, we argue for rethinking its role within more interactive and cognitively grounded learning frameworks 📘

While our findings suggest that Child-Directed Language (CDL) does not consistently enhance syntax learning in current models, they also highlight a key limitation: our models learn in ways fundamentally different from humans. 👶🤵‍♀️

We found that models trained on CDL consistently underperform compared to those trained on Wikipedia, regardless of language, model architecture, or syntactic evaluation task, even when controlling for lexical frequency effects 🔠

This work is developed in close collaboration with my amazing supervisors @arianna-bis.bsky.social @yevgenm.bsky.social and colleague @jumelet.bsky.social to whom I’m deeply grateful!

Child-Directed Language Does Not Consistently Boost Syntax Learning in Language Models Seminal work by Huebner et al. (2021) showed that language models (LMs) trained on English Child-Directed Language (CDL) can reach similar syntactic abilities as LMs trained on much larger amounts of ...

“Child-Directed Language Does Not Consistently Boost Syntax Learning in Language Models”

I’m happy to share that the preprint of my first PhD project is now online!

🎊 Paper: arxiv.org/abs/2505.23689