The power consumption of the device will be so small that it can collect energy directly from the surrounding environment. The project has received funding from the Villum Experimental Program.
In recent years, artificial intelligence (AI) has exploded, but despite tremendous progress, the capabilities required to run AI algorithms continue to grow.
In sharp contrast, the human brain only needs about 20W of power to perform more than 10 trillion (10,000,000,000,000,000) operations. This is 12 orders of magnitude better than modern supercomputer technology.
“This is why we are conducting in-depth research to develop new hardware that simulates the structure of the human brain, including neurons, synapses, and neural networks, called brain-inspired computing (BIC). However, even though we have managed to drastically reduce the cost of AI algorithms Power consumption, but in terms of size and energy efficiency, BIC is as efficient as the human brain, and there is still a long way to go,” said Assistant Professor Hooman Farkhani. In the Department of Electrical and Computer Engineering at Aarhus University.
He has just received a grant of 1.9 million Danish kroner from the Villum experimental program for a new project seeking to develop a nanoscale BIC system.
“If we succeed, we will have the first BIC system that does not exceed a particle of dust and has such low power consumption that it can collect energy directly from the surrounding environment. In other words, you will not need a power supply, which will A series of new AI applications that were not possible before paved the way,” Hooman Farkhani said.
The project, called Spin-Grain, is one of 51 projects that have just received a total of approximately 99 million Danish kroner grants from the Villum Experimental Program of the Villum Foundation.
The program donated money for “Bold Research Experiments” and commended Villum Kann Rasmussen, the founder of the Villum Foundation, and his tireless experimental methods for life.
The researchers behind the 51 experiments ranged from doctoral students to professors, representing a wide range of different nationalities. In addition to Aarhus University, scholarships were awarded to Aalborg University, Copenhagen University, Technical University of Denmark, University of Southern Denmark, IT University, Roskilde University and GEUS.