Call launch
Heterogeneous integration and neuromorphic computing technologies for future semiconductor components and systems
On 6 February 2024, the Chips JU launched a joint call with the Republic of Korea for low TRL actions on Heterogeneous integration and neuromorphic computing technologies for future semiconductor components and systems. It was a one phase call.
EU budget
Timeline
The call 2024-3 was a one phase call (no Project Outline phase) with the following schedule:
Call timeline
Submitted proposals
15 proposals were submitted for the 2024-3-RIA call.
Selected proposals
The projects are selected by the PAB based on the ranking of the experts, taking into account the overall consistency of the portfolio approach as well as national strategic priorities in accordance with article 12.1 of the SBA. Focus topics were selected first, and then the projects submitted from the bottom-up parts of the calls were selected. The results of the selection of proposals are summarized
Projects summary
NEHIL
Neuromorphic-Enhanced Heterogeneously-Integrated FMCW LiDAR
NEHIL aims to develop two innovative neuromorphic computing architectures for tackling the complex demands of modern data-intensive applications, one based on FeFET-based Compute-in-Memory (CIM) accelerators, and the other employing photonic integrated circuits based on reservoir computing (RC) principles, significantly easing manufacturing while enhancing the processing of dynamic data streams.
ViTFOX
Vision transformers with ferroelectric oxides
ViTFOX develops a ferroelectric-augmented intelligent semiconductors technology to demonstrate a Vision Transformer with energy efficiency better than 50 TOPS/W targeting AI-powered edge applications.
HAETAE
Heterogeneously integrated Multi- material Photonic Chiplets for Neuromorphic Photonic Transfer Learning AI Engines
HAETAE develops a multi-material PIC technology platform and aligns this along photonic Neural Network architectures capable of operating along the principles of Transfer Learning methods. This co-integrated PIC platform will provide a learning engine that can support one order of magnitude improvements along all critical performance metrics of AI chipsets
ENERGIZE
Energy-efficient Neuromorphic 2D Devices and Circuits for Edge AI Computing
ENERGIZE will implement novel neuromorphic hardware based on two-dimensional (2D) materials for energy efficient artificial intelligence (AI) through wafer-scale growth of 2D materials for neuromorphic devices, reliable fabrication and characterization processes of related devices, departing from conventional large energy consumption of von Neumann CMOS AI hardware.
