Beijing, 19 August 2024 (TDI): The Chinese Academy of Sciences’ Institute of Automation announced on Friday that a group of scientists from China has developed a unique brain-inspired network AI model.
The model is based on internal complexity to address issues with previous models, like excessive computational resource consumption.
To create more general-purpose AI that allows models to have a wider range of cognitive abilities is one of the main goals of the present artificial intelligence (AI) development effort.
According to Li Guoqi, an Institute of Automation researcher, as means of obtaining general intelligence based on “external complexity,” huge models currently employ a popular strategy that involves building larger, deeper, and wider neural networks based on the law of scale.
Unfortunately, this method has drawbacks like unsustainable energy and computer resource usage and interpretability issues.
However, the human brain contains approximately 1,000 trillion synaptic connections and 100 billion neurons, each of which has a unique intrinsic structure. However, a human brain uses just about 20 watts of power.
Also read: Chinese Firms Use Robotic Guardians To Combat AI Hackers
The researchers at the Institute of Automation and other research centers like Tsinghua University and Peking University employed an “internal complexity” strategy to develop general intelligence. Their inspiration came from the dynamics of brain neurons.
They conducted experiments that confirmed the internal complexity model’s dependability and efficacy in handling challenging tasks.
They also offered viable solutions for improving and optimizing the practical performance of AI models, as well as new approaches and theoretical support for incorporating the dynamic aspects of neuroscience into AI.