Master MVA
On this page, I will present the courses I took and link the github pages of some of the projects I did for them.
First semester
Learning courses
Convex optimization and applications in machine learning
I implemented a barrier method implementation in Python for quadratic optimization problems as an homework:
Graphs in machine learning
I did a project with Abel A. on re-implementing and studying in depth the paper “Online Influence Maximization with Local Observation” of Gábor Lugosi, Gergely Neu and Julia Olkhovskaya.
Online Influence Maximization with Local Observations
Reinforcement Learning
Data/Modelisation courses
Image denoising: the human-machine competition
Introduction to digital imaging (Imagerie numérique)
For this course I did a project with Paul B. on the paper “Style Transfer by Relaxed Optimal Transport and Self-Similarity” by Nicholas Kolkin, Jason Salavon and Greg Shakhnarovich. We used the existing implementation to push the analysis of the paper further, trying experimentations and extensions.
Style Transfer by Relaxed Optimal Transport and Self-Similarity
Object Recognition and Computer Vision
For this course I did a project with Hugo A. on the paper “SinGAN: Learning a Generative Model from a Single Natural Image” by. We used the existing implementation to create powerful denoising algorithms, furthering the capacities of algorithms such ad FFDNet.
SinGAN: Learning a Generative Model from a Single Natural Image
Second semester
Learning courses
Modelisation in Neurosciences
For this course, I did a project on the papers “Network Plasticity as Bayesian Inference” and “Synaptic Sampling: A Bayesian Approach to Neural Network Plasticity and Rewiring” by David Kappel, Stefan Habenschuss, Robert Legenstein and Wolfgang Maass. I studied these papers and reimplemented one network, a Restricted Boltzmann Machine, with Synaptic Sampling.
Synaptic Sampling: A Bayesian Approach to Neural Network Plasticity and Rewiring
Sequential Learning
Deep Learning in Practice
Kernel methods in machine learning
Data/Modelisation courses
Audio analysis
For this course, I implemented a CNN-based note classifier for music. // To be added