Prof. Patrick Rinke Awarded Academy Grant for Developing Biologically Inspired Computing Systems

Prof. Rinke’s three-year joint project with VTT aims to make demanding AI computing tasks use less power while maintaining performance.
Image of neuromorphic circuits
The AI4AI project aims to employ currently available AI tools to design new neuromorphic AI devices. Image: Patrick Rinke/Aalto University, Sayani Majumdar/VTT.

Professor Patrick Rinke of the Department of Applied Physics has been awarded an Academy of Finland grant for a project exploring the development of biology-inspired materials for improved sustainability in artificial intelligence (AI) computation. The project, entitled “AI software-based material design for sustainable AI hardware” (AI4AI), is a collaboration with Sayani Majumdar, Senior Scientist at VTT. VTT is in charge of coordinating the project.

In general, artificial intelligence tasks are extraordinarily demanding for conventional computers; the machines slow down and use enormous amounts of power. As demand for AI increases around the world, so does its environmental impact.

Inspired by the energy efficiency of the human brain, scientists have begun to develop hardware specifically for AI computation. Such neuromorphic hardware imitates the architecture of neuro-biological structures, but its development is still in the early stages. The three-year AI4AI project applies an AI-based material design strategy to create the next generation of neuromorphic AI hardware. The researchers develop electronic synapses and neurons with ultra-low power consumption that maintain fast computational speed, accuracy, and learning capacity in AI tasks.

“We use neural-networked-based AI methods on conventional computer chips to develop new devices that realize neural networks directly on a chip. Such neuromorphic AIs use a lot less power than conventional AIs on conventional chips and are therefore a lot better for the environment,” Professor Rinke says.

According to Rinke, the project is a fascinating opportunity to change the way we perceive AI.

“I really liked the idea of using AI to design the next generation of AI. We always think of AIs as living in the digital realm, but in this project, we are also changing the physical realm (i.e., computing hardware) that the AI runs on. Future AIs are therefore not only evolving digitally, but also physically,” Rinke says.

The funding was granted as part of the Academy of Finland’s call “ICT 2023: Sustainable and Energy-Efficient Solutions for Future ICT”. The AI4AI project starts in 2023.

Finnish Center for Artificial Intelligence

Finnish Center for Artificial Intelligence FCAI (external link)

International research hub for artificial intelligence

Photo showing researchers from CEST in front of a building.

Computational Electronic Structure Theory (CEST)

CEST is developing electronic structure and machine learning methods and applying them to computational materials science problems.

Department of Applied Physics
  • Published:
  • Updated:
URL copied!

Read more news

eit services
Cooperation Published:

EIT Digital Innovation Factory (Apply by 15th May 2023)

Are you planning to boost your early-stage digital deep tech startup?
Retuperän WBK:n Audio Dubbado, kuva: Atte Mäkinen.
Campus, Studies Published:

Audio Dubbado: Retuperän WBK represents firefighter-like dignity and good manners

'After I had played in the fire brigade for a few years, I thought that maybe it would be good to have some excuse to spend so much time in Otakaari. I decided it was time to apply to study at Aalto.'
Petteri Isotupa
Cooperation, Studies Published:

Aalto University alumnus Petteri Isotupa: ‘Dare to think and to ask why’

With his bequest to Aalto University, alumni Petteri Isotupa wants to support future generations’ opportunities for a good education.
Maarit Karppinen
Research & Art Published:

Significant funding for thin-film technology research - materials to be developed can be used, for example, in energy storage

The search for new, precisely tailored thin-film materials is accelerating for a wide range of next-generation applications in fields such as energy storage and IoT.