Scientists from the Institute of Automation under the Chinese Academy of Sciences in the Chinese mainland have unveiled a groundbreaking brain-inspired artificial intelligence (AI) model aimed at addressing the high energy consumption challenges inherent in traditional AI systems. Published in the journal Nature Computational Science, this research marks a significant shift from the conventional approach of scaling up neural networks to achieve general intelligence.
Researcher Li Guoqi highlighted that the current method, termed \"external complexity,\" involves building larger and more intricate neural networks, which leads to substantial energy and computational resource use while often lacking interpretability. In stark contrast, the human brain operates with approximately 100 billion neurons and 1,000 trillion synaptic connections on just 20 watts of power, demonstrating remarkable efficiency.
Inspired by the internal dynamics of the brain, the team from the Institute of Automation, in partnership with Tsinghua University and Peking University in the Chinese mainland, implemented an \"internal complexity\" approach to AI development. Their experiments have validated the model's capability to handle complex tasks efficiently, offering a novel pathway for incorporating neuroscience principles into AI advancements and optimizing performance.
Reference(s):
Chinese scientists develop brain-inspired AI to cut energy demand
cgtn.com
