PhD opportunities

Carbon-aware High Performance Artificial Intelligence

Thesis proposal

Area of expertiseInformatique temps réel, robotique et automatique - Fontainebleau
Doctoral SchoolISMME - Ingénierie des Systèmes, Matériaux, Mécanique, Énergétique
SupervisorTADONKI Claude
Co-supervisorMESRI Youssef
Research unitMathématiques et Systèmes
ContactClaude Tadonki
Keywordsmachine vision, fluid dynamics, deep learning
AbstractThe aim of this PhD is to foster an energy-aware design of AI technology. The focus will be given to two mainstream computing platforms: standard HPC solutions (clusters, and datacenters) and embedded devices. In both, the consumed energy is critical. In the former, the available energy is considered unlimited, thus the main concern in this aspect is generally on the associated cost (electricity and cooling system) and the carbon footprint [3]. In the latter, energy consumption is taken into account at the design time since the targeted devices are likely to be battery powered. Nonetheless, the battery life-cycle’s ecological burden is non-negligible too.
ProfileMaster in Computer science, with interest in applications (vision, physics).
FundingConcours pour un contrat doctoral