ERC Proof of Concept Grant for Erika Covi’s Brain-Inspired AI Hardware Project “FELICE”

Dr. Erika Covi, has received a prestigious European Research Council (ERC) Proof of Concept grant for her project FELICE – Ferroelectric Leaky Integration for Computational Efficiency.

FELICE — aptly named after the Italian word for happy — addresses a key challenge in artificial intelligence: the need for compact, low-power hardware capable of real-time learning and decision-making at the edge. Existing AI hardware is limited by high energy consumption and the separation of memory and processing, while many neuromorphic systems still rely on external training.

Covi’s project introduces a new neuromorphic architecture based on ferroelectric leaky integrate-and-fire (FeLIF) neurons, built from non-volatile FeFET memory technology. These components enable efficient on-chip learning and temporal data processing, while significantly reducing power and area requirements.

A key feature of FELICE is the BRUNO training methodology, which supports adaptable learning across different timescales — crucial for dynamic, edge-based AI applications. The system will be implemented at the 28 nm technology node, ensuring compatibility with current CMOS processes and making the technology scalable and commercially viable.

With this ERC grant, Covi and her team will validate the FELICE architecture as a next-generation solution for energy-efficient, adaptive AI hardware.

More information and a full list of awarded ERC PoC projects.