Chinese researchers have made a breakthrough in chip manufacturing technology with the development of low-power “brain-inspired” chips designed for artificial intelligence (AI) applications.
A collaborative effort led by the Chinese Academy of Sciences (CAS) is a neuromorphic chip capable of dynamic computing. By combining a dynamic visual sensor and a neuromorphic chip in a single chip, it consumes significantly less quiescent power.
According to Li Guoqi, a researcher at CAS’s Institute of Automation, the Spec uses only 0.7 milliwatts when performing visual tasks, resulting in impressive energy efficiency. .
The chip’s innovation lies in its system-level design, combining algorithms, software and hardware. This approach exploits “brain-inspired” computation, similar to higher-level brain functions such as dynamic attention allocation.
New research in the journal Nature Communications takes an important step toward increasing the efficiency of the human brain.
“The miracle of the human brain’s nervous system is that it only uses 20 watts of power, a fraction of what current AI systems require,” Lee said.
With computing demands leading to higher power consumption, Lee pointed to the potential of neuromorphic chips inspired by the structure and function of the brain.
The human brain can dynamically allocate attention based on stimuli known as attentional mechanisms. The researchers proposed “neuromorphic dynamic computing” using this principle to power the design of neuromorphic chips, thereby unlocking higher performance and energy efficiency, Lee said.
The research paves the way for the development of intelligent computing systems that are not only powerful but also incredibly energy efficient, Li added.
