Postdoc working on human cognition about abstract concepts at @behrenstimb's lab. Cog/comp-(neuro)scientist wannabe. Climber.
London, UK
Joined February 2021
- Tweets177
- Following180
- Followers493
- Likes254
Pinned Tweet
New paper just out in Cognitive Psychology: "A language of thought for the mental representation of geometric shapes" authors.elsevier.com/a/1g5%7…
Thanks to Kevin Ellis, Josh Tenenbaum, @StanDehaene and the reviewers!
See thread below for an outline👇
New preprint from the lab: "A language of thought for the mental representation of geometric shapes"
by Mathias Sablé-Meyer @MSableMeyer, with Kevin Ellis, Josh Tenenbaum and myself.
Find it here: psyarxiv.com/28mg4
Mathias Sablé-Meyer retweeted
“Where was I again?”
Our study published today nature.com/articles/s41586-0… reveals brain cells can form a coordinate system for our behaviours. Instead of locating where we are in the world, this coordinate system tells us “where we are” in a sequence of behaviours: 🧵below:
Mathias Sablé-Meyer retweeted
New ARC-AGI paper
@arcprize w/ fantastic collaborators @xu3kev @HuLillian39250 @ZennaTavares @evanthebouncy @BasisOrg
For few-shot learning: better to construct a symbolic hypothesis/program, or have a neural net do it all, ala in-context learning?
cs.cornell.edu/~ellisk/docum…
Help friends and colleagues better understand how humans process language! Good Old Fashion Psycho-Linguistics (GOFPL) with a fun hypothesis, but we promise there will be LLMs in the paper eventually 🤖
Participate: neurospin-data.cea.fr/exp/ja…
Share around to native English speakers!
Mathias Sablé-Meyer retweeted
1/🧵How early do children grasp mathematical patterns? Our latest study with @StanDehaene, @LubineauMarie and @MorfoisseTheo explores how young children’s drawings reveal their understanding of geometric patterns—from simple lines to complex curves like sinusoids and exponentials
Ahead of the event, brush up on Compositionality by watching replays of our online session from a month ago, featuring Akshita Srinivasan, @CognOriginsLab, @CiccioneLorenzo, @MorfoisseTheo and @PMuhleKarbe!
Find the links on workshop-compositionality-co…
Absolutely thrilled to share the program of our #CogSci2024 @cogsci_soc workshop 👉 doors open @ 8:30 👉 welcome @ 8:45 👉 talks++ @ 9:00 @LakeBrenden @liinapy AKJagadish @SbllDtrch @FoscaAR @GabyMohamady HWang @LiPingWangION @EltetoNoemi KEllis @MSableMeyer LWong @UoB_SoP 1/2
Mathias Sablé-Meyer retweeted
Absolutely thrilled to share the program of our #CogSci2024 @cogsci_soc workshop 👉 doors open @ 8:30 👉 welcome @ 8:45 👉 talks++ @ 9:00 @LakeBrenden @liinapy AKJagadish @SbllDtrch @FoscaAR @GabyMohamady HWang @LiPingWangION @EltetoNoemi KEllis @MSableMeyer LWong @UoB_SoP 1/2
Mathias Sablé-Meyer retweeted
#Workshop at #CogSci2024
Compositionality in Minds, Brains and Machines: A unifying goal that cuts across Cognitive Sciences
Organisers: @B_Pomiechowska, Rachel Dudley, Lio Wong, and @MSableMeyer
Visit workshop-compositionality-co… to learn about the program, speakers and more!
ALT Banner for a COGSCI2024 workshop titled 'Compositionality in Minds, Brains, and Machines: A Unifying Goal That Cuts Across Cognitive Sciences.' The workshop organizers are Barbara Pomiechowska, Rachel Dudley, Lio Wong, and Mathias Sablé-Meyer. The conference theme is 'Dynamics of Cognition,' held from July 24-27 in Rotterdam. The banner features a purple border and decorative geometric patterns on the right side.
Mathias Sablé-Meyer retweeted
How do we infer perceptual absences? Do beliefs about perception play a role in such inferences? And how does it all relate to cellist Natalia Gutman’s career decisions? Find answers in our new preprint with @moran_rani and @ClarePress:
osf.io/preprints/psyarxiv/gx…
A 🧵
Mathias Sablé-Meyer retweeted
Let’s talk *c o m p o s i t i o n a l i t y * 🧵
Before #CogSci2024 @cogsci_soc and our workshop on compositionality in brains, minds, and machines, join us next week on Wed 26 June for 👉a special online session Compositionality in Maths & Nonlinguistic Reasoning 👈 1/n
Mathias Sablé-Meyer retweeted
Come work with us! Looking for a 👉PhD student interested in finding out what makes human cognition special, studying symbolic thought, compositionality, behaviour &🧠@PMuhleKarbe Sarah Beck @UoB_SoP @TheCHBH Lab meetings in 🌳👇 reach out by 24/5 more in🧵Please RT! 1/n
Mathias Sablé-Meyer retweeted
Is a low-level associative learning strategy sufficient to encode sparse community network? In our new paper out in the Journal of Neuroscience, we argue that yes, see how in this thread! ▶️jneurosci.org/content/44/14/…
Every other lab meeting, when I'm confused thanks to my poor mouse neuroanatomy knowledge:
wait, the oiah alosias region is involved in fine grained sturbly control?
Mathias Sablé-Meyer retweeted
Happy to share my new first-author paper on visual sequence representation in the prefrontal cortex.
Thanks to my collaborators @fanispa @TvanKerkoerle @StanDehaene @bechirjarraya and @jo_bellet
cell.com/cell-reports/fullte…
Mathias Sablé-Meyer retweeted
New manuscript from the lab!
What is the neural code for reading?
Through simulations of convolutional neural networks, we propose a new hypothesis on how, in literate brains, neurons encode written letters and their relative positions within the word.
arxiv.org/abs/2403.06159
Mathias Sablé-Meyer retweeted
New paper from the lab on compositionality, just out in Cell Reports :
“Tracking the neural codes for words and phrases during semantic composition, working memory storage and retrieval”
Available in open access: cell.com/cell-reports/fullte…
A thread (1/n)
Mathias Sablé-Meyer retweeted
Test your math knowledge!
Our new experiment evaluates your knowledge of the math vocabulary in English. Click on the link below, the test only takes about 15 minutes.
Thank you for your participation!
neurospin-data.cea.fr/exp/ma…
Mathias Sablé-Meyer retweeted
New preprint (arxiv.org/abs/2309.17363) with incredible PNI second year co-first author, Declan Campbell! In this work, we show how ANNs can yield human behavior on abstract processing of geometric stimuli, providing an alternative to symbolic LoT models in this domain. [1/n]
