Titre:
Design Automation and Multi-Objective Optimizations for High-Performance Embedded Software
Conférencier:
Karol Desnos ,
INSA Rennes, France
Lieu:
Polytechnique Montréal, Pavillon Lassonde, L-2712 ,
Date et heure:
lundi le 11 novembre 2024 de
10:30 à 12:00
Résumé: This seminar summarizes a decade of research aimed at taming the design complexity of high-performance embedded systems, with contributions along three main research axes. The first research axis explored is the use of dataflow models of computation to program heterogeneous multicore chips. The second research axis explores the use of approximations as a tool to reduce the computational complexity of complex systems. The third research axis explored concerns the development of frugal artificial agents based on Tangled Program Graphs (TPGs).
Note biographique: Karol Desnos is an associate professor at the National Institute of Applied Science of Rennes, in France, and a researcher at the Institute of Electronics and Digital Technologies (IETR). His research interests primarily focus on the use of Models of Computation and automated Design Space Exploration techniques for the optimization of modern applications onto heterogeneous Multi-Processor Systems-On-Chips. Since 2019, he is leading the development of the open-source GEGELATI library, a lightweight framework for training and executing artificial intelligences based on tangled program graphs.