
19 European partners rally on next-gen neuromorphic memory
The program aims to develop process technology and hardware platforms based on emerging memory technologies for neuromorphic computing, addressing future applications in mobile devices that need complex machine-learning algorithms. Ultimately, the end-goal of this pan-European collaboration effort is to cram the capabilities of today’s cloud-based server racks into neuromorphic chips able to execute the same algorithms within battery-powered mobile devices such as cars and smartphones (at the edge of the internet-of-things).
In the coming years, the demand for edge artificial intelligence and machine-learning algorithms is only set to grow further, running increasingly complex computational algorithms for natural-language processing, face-recognition-based security systems or autonomous vehicles. Today, high-end server parks process the data in the cloud, with the inevitable data latency and energy inefficiencies associated with sending the data, let alone privacy issues. Hence, the ultimate edge artificial intelligence applications will require intelligent energy-efficient local processing.
“We are delighted to enter in such broad European collaboration effort on Edge Artificial Intelligence, gathering the relevant stakeholders in Europe, including CEA-Leti and Fraunhofer, two of our most renowned colleague research centers in Europe. Thanks to our combined expertise, we can scan more potential routes forward than what would be possible by each of us individually, and as such, position Europe in the driver seat for R&D on AI. Imec looks forward to the progress we can make together in the TEMPO project and hopes this will lead to more similar collaborations in the future. Behind the scenes, we are already defining more public and bilateral agreements with several of the partners involved” said imec’s CEO Luc Van den hove, in a company statement.
TEMPO will leverage the process technology platforms that are being developed by the European research technology organizations and cooperating foundries in the project, and combine it with the application and hardware knowledge from further partners. The TEMPO project will evaluate the current solutions at device, architecture and application level, and build and expand the technology roadmap for European AI hardware platforms. The project will leverage MRAM (imec), FeRAM (Fraunhofer) and RRAM (CEA-Leti) memory to implement both spiking neural network (SNN) and deep neural network (DNN) accelerators for 8 different use cases, ranging from consumer to automotive and medical applications.
“It is our aim to sweep technology options, covering emerging memories, and attempt to pair them with contemporary (DNN) and exploratory (SNN) neuromorphic computing paradigms. The process- and design-compatibility of each technology option will be assessed with respect to established integration practices and meet our industrial partner roadmaps and needs to prepare the future market of Edge IA where Europe is well positioned with multiple disruptive technologies”, expressed Emmanuel Sabonnadiere, CEO at CEA-Leti.
“A key enabler for machine learning and pattern recognition is the capability of the algorithms to browse through large datasets. Which, in terms of hardware, means having rapid access to large memory blocks. Therefore, one of the key focal areas of TEMPO are energy efficient nonvolatile emerging memory technologies and novel ways to design and process memory and processing blocks on chip”, said Prof. Hubert Lakner, Director of the Fraunhofer Institute for Photonic Microsystems (IPMS) and Chairman of the Board of Directors of the Fraunhofer Group Microelectronics.
TEMPO was kicked off on the 1st of April 2019 and has a duration of three years. The consortium of this ambitious project consists of no less than nineteen members. Imec takes the lead as the sole Belgian consortium partner. The other consortium members are, for France: CEA-LETI, ST-Microelectronics Crolles, ST-Microelectronics Grenoble, Thales Alenia Space and Valeo. For Germany: Bosch, Fraunhofer EMFT, Fraunhofer IIS, Fraunhofer IPMS, Infineon, Innosent, TU Dresden and Videantis. For the Netherlands: imec the Netherlands, Philips Electronics and Philips Medical Systems. For Switzerland: aiCTX and the University of Zürich.
TEMPO – www.ecsel-tempo.eu
Related articles:
Multi-bit memory chip stacks on top of CMOS for edge-AI
CEA-Leti proves RRAM-based TCAMs viable for neuromorphic processors
University College London spins out ReRAM startup
Spin-orbit torque MRAM devices switch field-free
Spin Memory: Using MRAM in place of SRAM
SiOx ReRAM selected for neuromorphic research
Applied, ARM to develop CeRAM for neuromorphic applications