Mathias Sablé-Meyer
Normalien, PhD at PSL/Collège de France & NeuroSpin
Present situation
Credit to Kaspar Ravel for this picture.
Starting a Ph.D under the supervision of Stanislas Dehaene at PSL/Collège de France, with a lab at NeuroSpin, CEA. Working title: “The cognitive representation of structured sequences: an exploration of humans’ outstanding capacity for abstraction through non representative geometry”.
Reach me: Academic email
Teaching
Highlights of the past situation
During the course of my higher education, I was a visiting researcher for half a year in the following places:
- MIT, Cambridge US in 2018, doing program induction with Josh Tenenbaum
- École Normale Supérieure, Paris, 2017 doing linguistics with Salvador Mascarenhas
- Oxford, UK in 2014 doing theoretical computer science with Luke Ong
Master:
- M2 CogMaster cohabilité ENS/EHESS/Paris Descartes
- M1 MPRI cohabilité ENS/ENS Cachan/Polytechnique/Paris Diderot
Past situation
Click to see more details about my academic path.
2017-2016 — Research @ MIT with Josh Tenenbaum
“Library Learning for Neurally-Guided Bayesian Program Induction”, one paper accepted, moving toward more cognitive science stuff, ongoing.
2017-2016 — Research @ ENS with Salvador Mascarenhas
Reasoners as mini-scientists: hypothesis testing and pragmatics. Two articles in prep. from this work.
2016-2017 — M2 @ CogMaster
Internship at NeuroSpin with Stanislas Dehaene, working on behavioural research and psychophysics of structured sequences in the visual modality. Academic report bellow, there is an article in prep..
2015-2016 — Gap year @ Mediterranean Sea
A dinner at Ligia, Greece with Félicien.
There’s a french blog about it. I went sailing and climbing for a whole year with a friend.
2014-2015 — M1 @ MPRI
Internship in Oxford, UK with Luke Ong, working on semantics of meta-programming and translations from/to other paradigms withing lambda-calculus. Academic report bellow.
2013-2014 — L3 @ ENS Cachan
Internship in INRIA, Lille with Romaric Gaudel, working on machine learning and cold-start versions of the multi-armed bandit problem. Academic report bellow.
Other occupations of my time
Climbing
Whenever I can, here and there, cliffs, rocks, indoor, boulder or just trees.
Tinkering
Playing around breaking stuff to understand how it works. A few selected examples:
- Project Ocarust or playing with FFIs and Ocaml/Rust
- Messing with DeepL, just for fun.
- Messing with GPT-2 117M, again.
Reading, traveling around, listening to some music
Publications
Under review
2019
- Mathias Sablé-Meyer, Salvador Mascarenhas. Assessing the role of matching bias in reasoning with disjunctions, in review. pdf
Peer reviewed articles
2018
- Kevin Ellis, Lucas Morales, Mathias Sablé-Meyer, Armando Solar-Lezama, Joshua B. Tenenbaum. Library Learning for Neurally-Guided Bayesian Program Induction. NIPS 2018. Spotlight.
Talks & Seminars
- Invited young researcher at Centre l’Oubradou, “Where Art, Science & Education connect”, 2018
- LPPRD seminar, joint talk with Salvador Mascarenhas, invited by Philip Koralis, 2018, handout.
- The Experimental Philosophy Group, 2017, handout
Academic Reports
Structural compression of visual input?
- Visual Sequence Primitives in Humans, 2017, internship report from my work in the NeuroSyntax team of the UniCog lab in NeuroSpin, under the supervision of Stanislas Dehaene.
- Toward an unstaging translation for an environment classifier based multi-staged language, 2015, internship report from work supervised by Luke Ong in Oxford’s Computer Science Department.
- Algorithmes de recommandation séquentielle, 2014, internship report with INRIA Lille’s team SequeL, supervised by Romaric Gaudel.
- I had fun with Jacquin’s algorithm for fractal compression of images in 2013, you can find some of that here
Misc & Non Serious
Spare time electronic stuff.
- I quite enjoy electronics from time to time. The documentation (non serious!) for a small project I realised during my time at NeuroSpin — reverse engineer a food distribution system to interface it with a computer through USB.
- A course from CogMaster was validated through a mini thesis. It’s about Mathematical Intuitions and Monsters (in French).
- Every once in a while I put something in this modest blog.
- More on the current background?