Scientists have successfully integrated machine learning with lab-grown human brain organoids, marking a significant advancement in computing. Inspired by the intricate three-dimensional biological network of the human brain, consisting of around 200 billion cells and trillions of nanoscale synapses, researchers aim to develop AI hardware that mirrors its structure and efficiency.
Led by Feng Guo, an associate professor of intelligent systems engineering at Indiana University Bloomington, the research team utilized cerebral organoids as a crucial component in a computing process. Traditional hardware inputs electrical data into the organoid, which then processes this information to generate an output. This innovative approach opens the door to the creation of "biocomputers" that boast enhanced power and energy efficiency compared to conventional silicon-based counterparts.
The technique at the core of this breakthrough is known as reservoir computing. In this system, the organoid acts as a reservoir that stores and reacts to incoming information. An algorithm is trained to identify changes within the organoid caused by various inputs and interpret these changes to generate outputs.
"We can encode the information—something like an image or audio information—into the temporal-spatial pattern of electrical stimulation," explained Guo. The algorithm then decodes the organoid's response to the patterned electrical stimuli, enabling it to perform computational tasks.
While these brain organoids are simpler than a full human brain, their ability to adapt and evolve in response to stimulation, akin to human learning, holds immense potential for revolutionizing AI. This pioneering work not only provides a glimpse into the development of biocomputers but also yields insights into the functioning of the human brain.
The research, published in the journal Nature Electronics, signifies a groundbreaking fusion of biology and technology, paving the way for a future where the two disciplines converge to unlock unprecedented possibilities in computing.