Description of the thesis topic

This thesis project focuses on the development of an innovative neuromorphic analog computer, capable of solving NP-hard combinatorial optimization problems, such as the travelling salesman problem or the “max-cut” problem. Using a platform of reconfigurable analog circuits of the FPAA (Field-Programmable Analog Array) type, this project proposes to overcome the limitations of conventional digital systems by exploiting non-linear dynamics inspired by the human brain. This analog computing approach offers significant advantages in terms of energy efficiency and convergence speed, to tackle complex problems encountered in various scientific fields. The PhD student will work as part of an interdisciplinary team with experts from Georgia Tech-CNRS, Georgia Tech Atlanta, and CentraleSupélec, while developing cutting-edge expertise in nonlinear dynamics and neuromorphic computing.

Work Context

The PhD student will work in the “Photonic and Terahertz Imaging” team of IRL 2958 Georgia Tech-CNRS, located at Georgia Tech Europe in Metz. This team specializes in nonlinear dynamics, terahertz imaging and analog computing. A key collaboration of the project will be with Prof. Jennifer Hasler of Georgia Tech Atlanta, a pioneer in the field of reconfigurable analog platforms (FPAA) and low-energy neuromorphic systems. The PhD student will benefit from direct access to her unique expertise, in particular for the implementation of the analog calculator on the FPAA platform. The PhD student will also interact regularly with internationally renowned researchers such as Prof. David Citrin and Dr. Alexandre Locquet. The project will offer a stimulating interdisciplinary research environment, with regular exchanges with experts in electronics, applied physics and artificial intelligence.

The position is located in a sector under the protection of scientific and technical potential (PPST), and therefore requires, in accordance with the regulations, that your arrival is authorized by the competent authority of the MESR.

Constraints and risks

The project includes technical challenges, such as optimizing the FPAA platform to solve NP-hard problems, which requires expertise in analog electronics. The performance of neuromorphic systems may be subject to unforeseen circumstances, and the international collaboration, in particular with Prof. Jennifer Hasler, will require good time management. The candidate will also need to pass ZRR (zone à régime restrictif) clearance.

Additional Information

Master's degree in engineering or science. Solid background in electronics, microelectronics or analog systems. Experience with analog or reconfigurable computing platforms (FPGA/FPAA). Knowledge of non-linear dynamics or complex systems desirable. Skills in combinatorial optimization and NP-hard problem solving desirable. Interest in artificial intelligence and neuromorphic computing. Ability to work independently in an interdisciplinary research environment. Very good command of the English.