In the quest to evolve artificial intelligence beyond mere algorithms into systems that emulate human-like intelligence, a groundbreaking concept has emerged from an international collaboration of scientists hailing from Dortmund, Loughborough, Kiev, and Nottingham. Their innovative approach, recently highlighted in Nature Communications, draws inspiration from the intricacies of human eyesight to potentially revolutionize future artificial intelligence systems.
At the core of their concept lies the creation of an on-chip phonon-magnon reservoir for neuromorphic computing. This reservoir, developed on a chip measuring a mere 25x100x1 cubic microns, harnesses the synergy of acoustic waves (phonons) and spin waves (magnons) to emulate the functionality of the human brain in processing complex external signals.
"The potential of the physical system proposed as a reservoir was immediately obvious for us because of its amazing combination of variability and multidimensionality," remarks Professor Alexander Balanov from Loughborough University, one of the visionaries behind the concept.
The concept mirrors the intricate workings of the human brain's sensory organs, which efficiently convert external stimuli into electrical signals processed through a vast network of neurons and synapses. This ability to process complex signals rapidly forms the basis for human responses to external stimuli.
"Our concept is very promising because it is based on the conversion of the income signal to high-frequency acoustic waves like in modern wireless communication devices," explains Dr. Alexey Scherbakov, who led the project at TU Dortmund University.
The chip's innovative design, incorporating a multimode acoustic waveguide and a patterned magnetic film, facilitates the pre-processing of information through conversion to a propagating phonon-magnon wavepacket. This results in a remarkable reduction of computational resources and training time, akin to the efficient signal processing observed in human vision.
"The functionality of the developed reservoir is based on the interference and mixture of the optically generated waves, which is very similar to the recently suggested mechanism of the information processing in the biological cortex," highlights Professor Sergey Savel'ev, emphasizing the resemblance of the concept to the functionality of the human brain.
This pioneering concept not only holds promise for advancing artificial intelligence but also hints at a future where compact and efficient neuromorphic computing systems, inspired by the complexity of human vision, could shape the next generation of intelligent technologies. As technology progresses, the prospect of mobile devices aiding in making decisions akin to human thought processes looms on the horizon, propelled by innovations like the on-chip phonon-magnon reservoir.
