Pinned Repositories
Anytime-strategies-for-best-arm-identification
Project for the MVA master to compare anytime sampling strategies in the MAB problem. In collaboration with Zineb Belkacemi.
CTS
Simulations for "Statistical Efficiency of Thompson Sampling for Combinatorial Semi-Bandits"
early_alignment
Code for the paper "Early alignment in two-layer networks training is a two-edged sword"
eboursier.github.io
Personal Page
GFdynamics
Code used for "Gradient flow dynamics of shallow ReLU networks for square loss and orthogonal inputs"
making_most_of_your_time
Making the most of your day: online learning for optimal allocation of time
MVA-NBA-Challenge
Data Challenge for MVA course by Stephane Mallat. Work with Alexandru Dragomir.
penalising_biases
Code used for the paper "Penalising the biases in norm regularisation enforces sparsity"
regularized_private_learning
Code for the paper 'Private Learning and Regularized Optimal Transport'.
sic-mmab
SIC-MMAB code.
eboursier's Repositories
eboursier/GFdynamics
Code used for "Gradient flow dynamics of shallow ReLU networks for square loss and orthogonal inputs"
eboursier/sic-mmab
SIC-MMAB code.
eboursier/Anytime-strategies-for-best-arm-identification
Project for the MVA master to compare anytime sampling strategies in the MAB problem. In collaboration with Zineb Belkacemi.
eboursier/penalising_biases
Code used for the paper "Penalising the biases in norm regularisation enforces sparsity"
eboursier/CTS
Simulations for "Statistical Efficiency of Thompson Sampling for Combinatorial Semi-Bandits"
eboursier/regularized_private_learning
Code for the paper 'Private Learning and Regularized Optimal Transport'.
eboursier/making_most_of_your_time
Making the most of your day: online learning for optimal allocation of time
eboursier/early_alignment
Code for the paper "Early alignment in two-layer networks training is a two-edged sword"
eboursier/eboursier.github.io
Personal Page
eboursier/MVA-NBA-Challenge
Data Challenge for MVA course by Stephane Mallat. Work with Alexandru Dragomir.