Zhejiang University in China has recently achieved a significant milestone by developing the world’s largest artificial brain simulation supercomputer, named Darwin Monkey. This advanced supercomputer features more than 2 billion neurons and 100 billion synaptic connections, mirroring the neural architecture of a macaque brain. As a result, it is capable of performing complex cognitive functions, including visual and auditory processing, language comprehension, and mathematical reasoning.
According to reports from the South China Morning Post, the Darwin Monkey supercomputer is equipped with 960 specialized neuromorphic chips known as Darwin 3. A representative from the laboratory stated that this innovative computer system is a substantial step toward realizing advanced intelligence that resembles human brain function.
One of the standout features of the Darwin Monkey is its remarkably low power consumption, operating at only about 2,000 watts under normal conditions. The neuromorphic computing technology employed in this supercomputer mimics the structure and processes of neurons and synapses in the human brain, distinguishing it from traditional deep learning techniques. This advancement allows for the deployment of artificial intelligence (AI) in a manner that is parallel, flexible, and energy-efficient.
Professor Pan Gang, director of the National Center for Brain-Computer Intelligence at Zhejiang University, highlighted that the capabilities of Darwin Monkey, including its parallel processing, large-scale operations, and low power requirements, are set to establish a new paradigm in the computing landscape.
Furthermore, the Darwin Monkey system is seen as a foundational technology that could replace biological experiments, thus advancing brain science research. It has the potential to accurately simulate the brains of various animals, including monkeys, mice, and zebras.
Before the development of Darwin Monkey, Intel introduced the Hala Point supercomputer, which contained approximately 1.15 billion neurons, significantly fewer than the 1.5 billion neurons in Darwin Monkey. Experts regard the Darwin Monkey as a prime example of implementing a biological brain-like computing structure, marking a pivotal advancement in both artificial intelligence and brain science.