Christophe Pallier

All NeuroImage and NeuroImage:Reports editors have resigned over the high publication fee, and are starting a new non-profit journal imaging-neuroscience.org This comes with great regret, and a huge amount of thought and discussion- please read announcement to get more details.

Happy to share our 7T fMRI paper about reading in bilinguals, with @chrplr , @aakashagr92 , @StanDehaene , and Laurent Cohen. science.org/doi/10.1126/sc…

Large language models change everything for linguistics, starting with Chomsky. Featuring: LLMs as scientific theories, response to prior takes, "why" questions in language, acquisition... and how the field should have seen this coming. Paper is here: lingbuzz.net/lingbuzz/007180

Predictions in brains and large language models: Our latest and is out at Nature Human Behavior: go.nature.com/3SKb3gX By, once again, our wonderful team @c_caucheteux and @agramfort

Joint work with @lakretz,@BertrandThirion,@chrplr. Thanks to @nilearn, @InriaMind, @NeuroSpin_91, @UNICOG1, @Inria_Saclay, @huggingface, @sklearn_inria and the open-source and open-science communities for making this possible!

Glad to share our #ICML2022 paper “Neural Language Models are not Born Equal to Fit Brain Data, but Training Helps”. In this work, we investigated the factors driving NLM's ability to fit fMRI brain data. With @lakretz, John Hale, @BertrandThirion, @chrplr arxiv.org/abs/2207.03380

Everyone seems to think it's absurd that large language models (or something similar) could show anything like human intelligence and meaning. But it doesn’t seem so crazy to me. Here's a dissenting 🧵 from cognitive science.

The efficiency put by institutions in breaking sci-hub seems to dwarf that put in funding science. It reveals how morally bankrupt the institutional elite is. Protecting interests leeching off science is more important than science. (I am failing to find a working sci-hub URL)

Il n’avaient le droit qu’à une prise.

I could write jokes for 800 years and I'd never think of something funnier than Trump booking the Four Seasons for his big presser, and it turning out to be the Four Seasons Total Landscaping parking lot between a dildo store and a crematorium.

Many journalists and political leaders in France explain that the pandemic is growing faster than expected. Here are the daily new cases in France since May 11th (end of the lockdown), log scale. How could it be more predictable? Where is the surprise?

MS Excel data files exceeding the maximum size prevents COVID-19 tracing in UK ...

J'avais été prodigieusement irrité par un "cours" d'épistémologie de Monsieur Raoult qui était d'une inculture crasse. Merci à @CovaFlorian de remettre les pendules à l'heure !

@sonjoachim Je vois pas le problème. Ils sont protégés par l'appli StopCovid, non ? ;-)

@EChabriere "moins il y besoin de puissance stat" -> NON. La puissance stat d'un test se calcule *a priori* avec une estimation de la taille d'effet. Plus celle ci est grande, plus l'effectif nécessaire pour atteindre une certaine puissance est petit. On ne baisse pas la puissance stat!

Puissance statistique et effectif Plus un effet est important moins il y a besoin de puissance statistique pour le prouver txrating.org/spc/polycop/Pu….

Moi quand j'essaie de tuer le moustique qui est dans ma chambre depuis 2h

@FranckRamus @StanDehaene "Priceless picture" indeed: Emmanuel wore a tie! Maybe it is a fake after all ;-).

@StanDehaene And that group picture is priceless! When and where was it taken? My memory is failing me.

My mentor Jacques Mehler died yesterday at the age of 83. He was an extraordinarily creative, fun and flamboyant cognitive psychologist, a pioneer of infant cognition, the founder of the journal Cognition. He brilliantly trained dozens of cognitive scientists in France and abroad

DEMAIN DANS LE CANARD Exclusif : François de Rugy justifie sa démission « ça suffit, j’en ai crustacé! »

Two new sites on-line: lexique.org and chrplr.github.io/openlexicon