Zhejiang University (ZJU)
The Darwin or “Wukong” monkey has over 2 billion artificial neurons and over 100 billion synapses – in an architecture very similar to the neural structure of a monkey.
Unlike traditional artificial neuronal networks, which follow classic computing principles and process data through continuously variable binary values, neuromorphic systems such as Darwin monkey are moved by neuronal networks with peaks.
As explained by, the also called, in English, spiking neural networks (SNNs) They imitate the way signs are transmitted between neurons in mammalian brains, responding to electrical signals to process and transmit data through connected and off (or peak) explosions of activity.
One Biological neurónia fires an electric impulse When the signs they receive from other neurons reach a strong enough level to trigger an answer. Artificial neurons in SNNs mimic this mechanismfiring only when they accumulated sufficient electrical input.
While software -based neuronal networks are a collection of automatic learning algorithms willing to emulate the human brain, SNNS physically replicate the way information moves between biological neurons.
This configuration allows SNNs to process data in parallel, potentially making it more powerful than conventional supercomputer architectures.
It can also be more efficient in energetic terms: Artificial neurons enter a brief rest period after each peak, during which they cannot respond to new entries. This limits the frequency with which they shoot, helping to reduce total energy consumption.
Num, the team behind monkey Darwin said the platform had already demonstrated its capabilities in cognitive tasks such as logical reasoning, content generation, and mathematical problem solving, using an AI model developed by the Chinese AI startup.
The Darwin monkey was designed by researchers at Zhejiang University and the Zhejiang Laboratory in China, who expect the new machine to serve as a simulation tool for neuroscientists, while providing a transition stage towards General Artificial Intelligence (AGI) – an artificial intelligence system (AI) that has intelligence and reasoning similar to humans.
The system is also being used to simulate animal brains with different levels of neural complexity, including unlucky fish and rats.
The previous holder of the neuromorphic computing record was Intel’s Hala Point system, which has 1.15 billion artificial neurons and 128 billion artificial synapses distributed by 140,544 processing nuclei.
