Shulou(Shulou.com)11/24 Report--

Thanks to CTOnews.com netizens for the clue delivery of the holy Buddha! This article comes from the official account of Wechat: SF Chinese (ID:kexuejiaodian), author: SF

Does artificial intelligence really have "intelligence"? Although artificial intelligence has made great progress so far this year, it may still be a far cry from what many people think. In our imagination, artificial intelligence should be like the "mechanical master" in the Terminator and the "mechanical master" in the Matrix, with the same "intelligence" as or even more than human beings. The ideal is very plump, the reality is very bony. Although it is called artificial intelligence, the current artificial intelligence is far from the level of real human intelligence.

(by Mei Lin / tr. by Robert Taylor)

In fact, our artificial intelligence does not have a "brain". It knows nothing, does not think, does not perceive, and even cannot understand the meaning of words and images. At present, artificial intelligence is trained with a large amount of data, and each training, the neural network will adjust the data received and output by each layer of unit in order to output more accurate answers. What artificial intelligence does is based on a large number of training data, multi-data processing, giving an answer similar to human "thinking".

For example, ChatGPT doesn't have any knowledge of geography, but it can tell you where the capital of France is. Of course, the premise is that during the training process, the input data includes the words "France", "capital" and "Paris". In answering this question, it only analyzes the relationship between words-France, the capital and Paris often appear together.

To this end, scientists imagine that it is possible to simulate the mechanism of the human brain in processing information and develop computing chips and corresponding hardware systems, so that computers and artificial intelligence can think like human beings and become "artificial intelligence" in the real sense.

To make artificial intelligence more like the human brain recently, a research team from the University of Science and Technology of China has prepared brain-like neuron devices based on vanadium dioxide phase change films, which are often called brain-like chips. Scientists at the Chinese University of Science and Technology used this new chip to simulate the dynamic connections of neuronal synapses under external stimuli, showing the process of multi-channel signal transmission and processing in brain-like neural systems.

If these brain-like chips are connected into artificial neural networks in the future, they will simulate the work of neural networks in the human brain, and computers and artificial intelligence equipped with such neural networks will think like humans.

What is a brain-like chip? Draw lessons from the information processing way of the human brain, use devices to simulate the nervous system of the brain, create a virtual brain, or merge the machine with life through brain-computer interaction, and create a super brain that combines the virtual brain with the biological brain. The ultimate goal is to build an artificial super brain, which is brain-like research. In the brain-like research, the neural device used to simulate the human brain neural system is the brain-like chip. To put it simply, brain-like chips are equivalent to the neurons in our brain, which make up the nervous system similar to the human brain.

Knowing what a brain-like chip is, some partners may ask what is the use of scientists using brain-like chips to simulate the human brain. At present, the results of brain-like research have been applied to multimodal target tracking, robot spatial localization, multi-scale brain simulation, robot group control and so on.

However, with the development of brain-like research and the increasing complexity of application scenarios, how to make computers and artificial intelligence more intelligent and closer to the human brain has become the focus of scientists. For this reason, scientists put forward the scheme of pulse neural network. compared with the current mainstream brain-like research schemes, the working mechanism of pulse neural network is closer to the human brain, so it has more advantages in efficiency, computation and application. Brain-like chip is the basis of constructing pulsed neural network.

Break through the bottleneck and simulate the synaptic connections of the human brain as early as 2005, the United States put forward a plan for brain-like research. Now, China, Japan and EU countries have also carried out brain-like research one after another. Although the brain-like research based on brain-like chips has become the current research hotspot, and scientists have made some achievements, there is still a long way to go before the large-scale commercial operation of brain-like chips.

Because the material for making chips is mainly silicon, the current artificial intelligence is called silicon-based artificial intelligence. It is difficult to use such silicon-based chips to simulate the nervous system of the human brain, especially synaptic connections.

The new brain-like chip developed by the research team of China University of Science and Technology uses alum dioxide phase change film as the basic material. Due to the reversible electrical conductivity transformation between insulators and metals at close to room temperature, alum dioxide is an ideal material for brain-like chips to simulate synaptic connections. In this study, scientists used a new method to break through the production bottleneck and directly simulate the process of synaptic connections between neurons in the human brain. The study was published in the October 4, 2023 issue of the journal SCIENCE ADVANCES.

References:

Quantum imaging of the reconfigurable VO2 synaptic electronics for neuromorphic computing | Science Advances

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.