CENTRALESUPELEC

Their core work

CentraleSupélec is a top-tier French engineering grande école located south of Paris, specializing in applied mathematics, signal processing, wireless communications, and machine learning. Their H2020 portfolio reveals deep strength in the mathematical foundations of AI — information theory, optimization, and deep learning — applied to domains ranging from 5G networks to combustion science. They contribute advanced modeling and algorithmic expertise to large European consortia, frequently as a third-party research contributor embedded within broader institutional partnerships. Their recent work shows expansion into quantum-HPC hybrid computing and plasma-assisted combustion, reflecting a pivot toward high-impact physics-based engineering challenges.

What they specialise in

How they've shifted over time

In their early H2020 period (2015–2018), CentraleSupélec focused heavily on the mathematical foundations of machine learning and wireless communications — information theory, deep learning architectures, and 5G radio technologies like massive MIMO. From 2019 onward, their work shifted toward applied industrial domains: cyber-physical systems, numerical engineering simulations, quantum computing infrastructure, and plasma-assisted combustion. This evolution shows a classic trajectory from theoretical AI/telecom research toward real-world engineering applications where those mathematical tools are deployed.

CentraleSupélec is moving from foundational ML/telecom theory toward high-value applied engineering — quantum computing, plasma combustion, and industrial cyber-physical systems — making them an increasingly attractive partner for hardware-intensive, physics-driven projects.

How they like to work

CentraleSupélec predominantly operates as a specialist contributor rather than a consortium leader — 9 of their 20 projects are as a third party (linked through a parent institution), with only 4 as coordinator. Their coordinated projects tend to be smaller ERC-scale grants (EUR 150K–170K), while their largest grant (GREENBLUE, EUR 2.5M) is a notable exception. With 243 unique partners across 28 countries, they function as a well-connected research node that brings mathematical and algorithmic depth to large European consortia without seeking to lead them.

CentraleSupélec has collaborated with 243 distinct partners across 28 countries, indicating a broad pan-European network. Their frequent third-party role suggests strong institutional ties (likely through Université Paris-Saclay) that channel them into major consortia across multiple sectors.

What sets them apart

CentraleSupélec offers a rare combination: rigorous mathematical and algorithmic expertise (optimization, information theory, deep learning) paired with applied engineering domains (combustion, wireless, cyber-physical systems). Unlike pure CS departments, they can bridge the gap between abstract AI methods and physical-world engineering problems. For consortium builders, they are the partner who brings the math that makes complex simulations and AI-driven control systems actually work.

Highlights from their portfolio

• GREENBLUE
  Their largest grant (EUR 2.5M) and an ERC Advanced Grant on plasma-assisted combustion for pollution reduction — signals a major institutional bet on clean energy research.
• STRUDEL
  Self-coordinated ERC Proof-of-Concept bridging information theory with deep learning — represents their core intellectual identity at the intersection of math and AI.
• WINDMILL
  A Marie Curie training network integrating wireless engineering with machine learning across 5G technologies — perfectly captures their dual telecom/AI expertise.