Page 20 - Study At Centrale Supélec 2021

Physics & Nanotechnologies/

Photonics

application to different fields : ultra-fast optics, telecommunications, all optical
ENGINEERING CHALLENGE TERM information processing.

CORE COURSE : ELECTIVE SERIES 2.5 (2,5 ECTS)
*ONLY TAUGHT IN FRENCH ON METZ CAMPUS
• CHAOS, FRACTAL AND COMPLEXITY
This lecture will provide students with fundamental tools and the framework
CONCENTRATION of the science of complexity. It will be illustrated by multiple examples from
recent research articles with applications to the field of engineering. Students
• NEURO-INSPIRED PHYSICAL SYSTEMS FOR will use analytical and numerical techniques for the resolution of problems.
INFORMATION PROCESSING (7,5 ECTS)
ELECTIVE SERIES 2.6 (2,5 ECTS)
Concentration course : Optimisation for neuro-inspired
computing with physical architecture • BIG DATA : COLLECTION, STORAGE AND ANALYSIS OF
DATA ON CLUSTERS AND CLOUD
The ever-increasing amount of data generated requires the development This course will investigate data engineering solutions that have appeared to
of novel processing/ analysing strategies. An alternate solution to Machine store and request large amount of data. It will also introduce algorithms and
Learning classification techniques consists in designing physical (hardware) mechanisms to filter and enhance large data sources and to select some
neuro-inspired architecture that allows the alleviation of parts, if not all, of subsets of data. Finally, it will introduce Machine Learning algorithms used to
these technological challenges. process and analyse selected subsets of data.
In this particular context and through the framework of reservoir computing,
this module offers to the students to discover the principle of conception and INTENSIVE SEMINAR
design of physical (hardware) neuronal networks. All along, the students will
METZ be familiarized with various optimisation techniques such as ridge regression *ONLY TAUGHT IN FRENCH ON METZ CAMPUS
or gradient descent and their accelerated versions, or stochastic heuristics (ex.
simulated annealing, genetic algorithms...).
Concentration project : Classification of image and
video signals with power-efficient photonic systems SEMESTER-LONG COURSES

Various software-based techniques exist to solve classification tasks and
SEMESTER BY COURSEWORK
they run on central processing units (CPU) or graphical processing units ENGINEERING SKILLS WORKSHOP (1 ECTS)
(GPU). The downside is usually reduced processing speed and high-power (MANDATORY)
consumption during learning and testing phase. These are strong motivations
for the development of alternative hardware (physical)-based architectures The objective is to explore and develop the skills expected from an engineer :
using analogue electronics or photonics. teamwork, project management, communication, complex problem solving,
The objective of this project will be : (i) the study of the ability of a photonic creativity, leadership, ethics, multicultural awareness.
architecture to classify video or images from publicly available databases, (ii)
the optimisation of its performance to be competitive with state-of-the-art PROFESSIONAL PRACTISE WORKSHOP (0,5 ECTS)
software solutions and (iii) provide an estimation of the power efficiency of (MANDATORY)
the architecture.
This course aims to continue supporting students in the discovery of the
ACADEMIC TERM engineering profession and in the construction of their professional project.

ELECTIVE SERIES 2.4 (2,5 ECTS)

• SMART PHOTONICS SYSTEMS
In the context of new technologies, light is increasingly used as a support for
calculating, transporting or storing information. The objective of this course
is to present recent developments in “ photonics ”, including lasers and their

20