AI and LLM in education
In my new role as scientific advisor at the French Ministry of Education, I have recently been looking for meta-analyses and RCTs about impact of AI (and LLMs) in education.
Latest posts
See also all posts ordered by category or date.
Meilleur developpeur de France 2023
Le 9 mars 2023 avait lieu la compétition Meilleur Développeur de France 2023. Les algorithmes nécessaires sont résoudre les problèmes sont assez simples, mais il faut coder rapidement. Après coup, et à tête reposée, voici comment on aurait pu résoudre certains des problèmes.
SWERC 2022 Practice Session - Bloggers
Abstract: Maintain two tables $t_0$, $t_1$. Updates: given $i,j,c$ increment $t_c$ between indices $i$ and $j$. After every update output $\sum_k \max{t_0[k],t_1[k]}$.
Deep Reinforcement Learning
Stable Baselines rely on TF 1.x but Stable Baselines v3 rely on PyTorch.
AlphaGo and Alpha Zero
In December 2016 we presented it at ENS Paris-Saclay with Étienne Simon but apparently I never wrote a blog post in the end.
Six dates importantes en algorithmique
- -300
- Algorithme d’Euclide (qui date peut-être de l’école de Pythagore 530 av. J.C.)
Quadrangle Inequality trick for dynamic programs
Application: Given an ordered list of keys with frequencies, build a binary search tree on those keys which minimizes the average query cost.
Pareto optimality
Compute the pareto set of a given set of points in 2 or 3 dimensions.
Multiplying polynomials
You are given two polynomials $P$ and $Q$ and want to compute their product. The polynomials are given in form of an array with their coefficients.