The core backbone of the center for excellence in brain science and intelligent technology in the Chinese Academy of Sciences, Professor Bi Guoqiang of University of Science & Technology China, believes that the core difficulty of brain intelligent research and development is that we are not enough to understand the structure and functional principles of the brain.
The human brain weighs about 1.4 kilograms, and the cerebral cortex has hundreds of billions of neurons, each containing tens of thousands of branches, forming a large and fine neural network. The brain is dealing with information through such a large scale neural network system, but the network’s wiring diagram is very complex, and there are many different types of neurons and synapses. It is necessary to imagine a comprehensive and complete line map with modern technology.
“At this stage, we can start building a simplified brain like model without fully understanding the principles of the brain, so as to achieve some” intelligent functions “. Bi Guoqiang, the present artificial neural network model, includes a deep neural network that mimics some of the most basic features of the neural network, and has achieved great success in dealing with the problem of classification recognition, but these “simple” networks have fundamental limitations on efficiency, power consumption, and generality. Ways to produce true intelligence.
“A focus at this stage is the development and application of new technologies, including machine learning (or brain like intelligence) technology, such as the existing artificial neural networks, to promote biological research on the brain network structure and learning rules, to accumulate a large amount of data and to understand the principles.” Bi Guoqiang said.
At the same time, by developing new software and hardware technologies, integrating new brain structures and working principles, the ability to improve brain intelligence technology is tried, which in turn promotes brain research. Through such a positive feedback iteration process, perhaps we can achieve the next breakthrough in the future.