Mohamed El Amine Seddik
Ph.D student in Random Matrix Theory and Machine Learning
mohamedelamine.seddik_at_cea.fr

About Me

Welcome to my personal webpage! I'm a french PhD student working on Random Matrix Theory and Machine Learning at CEA (Alternative Energies and Atomic Energy Commission). I am also attached to the L2S (Laboratory of Signals and Systems) laboratory of Centrale-Supélec (part of the University of Paris-Saclay). I work under the supervision of Mohamed Tamaazousti (CEA) and Romain Couillet (Centrale-Supélec).



My Thesis

The bigdata era has driven the recent development of new algorithms and methods, often based on elementary principles allowing to handle large amounts of data. However, these large dimensional data impair the behavior of traditional methods that deserve to be revisited under the eye of more elaborate tools and methods. A better understanding of these methods in the bigdata regime indeed induces possibilities of improvements, thereby leading to the development of more efficient algorithms. The Random Matrix framework provides a powerful tool to understand and analyse the behavior of simple data models (such as the mixture of Gaussians model) in the large dimensional setting, which is naturally the case in the BigData paradigm. My PhD thesis will aim at going beyond the simple models hypothesis, to develop new methods that are more appropriate to practical datasets (structured data, images, etc.).
Keywords: Random-Matrix-Theory, Machine-Learning, Deep-Learning, Image-Analysis.



Publications

2018

A Kernel Random Matrix-Based Approach for Sparse PCA


Mohamed El Amine Seddik, Mohamed Tamaazousti, and Romain Couillet
ICLR 2019 (submitted)

PDF BibTex

Learning More Universal Representations for Transfer-Learning


Youssef Tamaazousti, Hervé Le Borgne, Céline Hudelot, Mohamed El Amine Seddik and Mohamed Tamaazousti
PAMI 2018 Journal of IEEE Transactions on Pattern Analysis and Machine Intelligence (submitted)
Journal - Impact Factor: 8.32
PDF BibTex Code (Github)

From Outage Probability to ALOHA MAC Layer Performance Analysis in Distributed WSNs


Mohamed El Amine Seddik, Viktor Toldov, Laurent Clavier, and Nathalie Mitton
WCNC 2018 International Conference on IEEE Wireless Communications and Networking, Barcelona, Spain, April 2018
Oral
PDF BibTex

Teaching

Practical Signal-Processing at Centrale-Supélec


Signal Processing tutorials for first grader engineering students.
Temporal Representation, Frequency Representation, Fourrier Transform, Signal Filtering, Sampling, Audio and Image Signals Analysis etc.
[Matlab]

With Elisabeth Lahalle
Centrale-Supélec - 2018

Reviewing Activities (Participation)

CVPR 2018


IEEE Computer Vision and Pattern Recognition 2018: Two papers.




ICLR 2018


International Conference on Learning Representations 2018: Two papers.




NIPS 2017


Advances in Neural Information Processing Systems 2017: Two papers.




Projects

Variance Reduction in Monte Carlo Methods


Bibliography on Monte Carlo methods, variance reduction using functional control variables and reduction of dimension for the multi-dimensional case.
[Linear Algebra / Probability and Statistics / Matlab]

Mohamed El Amine Seddik under the supervision of François Portier (Télécom ParisTech)
Scholar Project 2017

Human Activities Classifation using Kernel Methods


Application of machine learning methods (in particular, kernel methods) to the problem of physical humain activities classification, using heterogeneous data from position sensors.
[Python / Scikit-Learn]

Mohamed El Amine Seddik (Télécom ParisTech)
Scholar Project 2017

Multi-Target Tracking


Bibliography and implementation of a multi-target tracking algorithm, silhouette detection, SVM classification and identification of targets.
[Python / Matlab]

Mohamed El Amine Seddik under the supervision of Rita Noumeir (ETS Montréal)
Scholar Project 2016

Neural Networks


Implementation of a document prediction and classification algorithm based on neural networks.
[Python]

Mohamed El Amine Seddik (ETS Montréal)
Scholar Project 2016