Putting a server in your pocket

Imec and 19 other research and industrial partners are combining to enable applications that now need servers to be performed by mobile devices at the edge.

The project will investigate the emerging memory technologies which will be needed for neuromorphic computing.

Called TEMPO, the three year initiative aims to deliver mobile devices that run complex machine-learning algorithms.

Today, applications such as smart home assistants with natural-language processing, face-recognition-based security systems or autonomous vehicles are processed by data centres in the cloud.

Sending data to the cloud costs energy, latency, and is often not preferred for privacy reasons.

As such, the ultimate edge AI applications require intelligent energy-efficient local processing.

TEMPO 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.

“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,” says Imec CEO Luc van den hove (pictured), “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.”