news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Editor of Every Economic Report: Du Yu

According to Science and Technology Daily, although artificial intelligence has surpassed humans in some aspects, this does not mean that it is really smart. On the contrary, it is still "very stupid and naive" in many cases and still needs to learn from the human brain.On August 16, Nature Computational Science published an important advance in the field of brain-like computing. Drawing on the complex dynamics of brain neurons, the team of researchers Li Guoqi and Xu Bo from the Institute of Automation, Chinese Academy of Sciences, together with researchers from Tsinghua University and Peking University, proposed a new method for building a brain-like neuron model.

"This research result improves the problem of increased computing resource consumption caused by the outward expansion of traditional models, and provides a new case for the effective use of neuroscience to develop artificial intelligence." said Li Guoqi, co-corresponding author of the paper.

More importantly, the model makes more efficient use of computing resources, while also significantly reducing the use of memory and computing time, thereby improving overall computational efficiency.

Li Guoqi said that this research provides new methods and theoretical support for integrating the complex dynamic characteristics of neuroscience into artificial intelligence and building a bridge between artificial intelligence and neuroscience. It also provides a feasible solution for optimizing artificial intelligence models and improving performance in practical applications.

According to Xinhua News Agency,Brain-like intelligence is also known as neuromorphic computingIt imitates the way the human brain works, allowing computer software and hardware to process information efficiently. Compared with traditional artificial intelligence, it has the characteristics of low power consumption and high computing power.

Neuroscience research has found that the adjustability of the strength of synaptic connections between neurons is one of the foundations of the brain's learning and memory functions. Changes in the strength of synaptic connections caused by past experiences can affect brain function.

Changes in the strength of synaptic connections, also known as synaptic plasticity, can enhance or inhibit neuronal activity, and their duration can range from a few milliseconds to hours, days or even longer.

According to Guangming Daily, if we can learn from the principle of synaptic plasticity, use some means to imitate and realize it, build artificial synapses similar to neural synapses, and then further build a system, we can better understand and simulate the way the brain works, further promote the cross-development of informatics and neuroscience, and realize brain-like computing.

"The human brain is the most complex information processing system discovered so far. It is unparalleled in its simplicity and efficiency. Therefore, experts in the field of artificial intelligence are wondering whether it is possible to develop more powerful artificial intelligence based on the brain." Speaking of brain-like intelligence, Wu Jingzhu, a professor at the School of Computer and Artificial Intelligence of Beijing Technology and Business University, previously told Science and Technology Daily.

In 1956, at the Dartmouth Conference where many computer science experts gathered, scientists proposed that it might be possible to establish a multidisciplinary collaborative working mechanism based on the two basic fields of neuroscience and cognitive science to develop artificial intelligence that reaches or even exceeds human levels.

Wu Jingzhu emphasized that brain science and cognitive science are the most important basic disciplines for developing brain-like intelligence. In recent years, with the development of imaging technologies such as functional magnetic resonance imaging, human cognition of the brain has greatly improved, which provides the necessary conditions for designing computer hardware and software based on the brain.

Han Liqun, a professor at Beijing Technology and Business University and an academician of the Academy of Engineering and Technology for the Developing World, believes that, in simple terms, the paths to achieve brain-like intelligence can be roughly divided into two categories: soft brain-like and hard brain-like. Wu Jingzhu explained that the main difference between the two lies in their different focuses, with the former focusing on algorithms and the latter on hardware. Although the paths are different, overall the two complement each other.

Hard brain-like technology mainly focuses on seeking breakthroughs in hardware materials. By developing neuromorphic chips (such as brain-like chips) and other media, based on disciplines such as bioelectronics and neuromorphic engineering, it simulates biological neurons and even the entire brain.Han Liqun said that the path of hard brain-like is to "first pursue similarity in form, then consider similarity in spirit." An ideal brain-like chip contains many processors equivalent to neurons, and the communication system between these processors is equivalent to nerve fibers. Structures such as synapses may also be simulated.

In the industry, companies such as Baidu, iFLYTEK, Alibaba, and Huawei have proposed some concepts related to brain-like intelligent applications in recent years. With the progress of brain-like scientific research, the "electronic brain" is turning from a textual concept into real-life application.

It is understood that the "Wentian I" brain-like computer, which has been officially put into use, has an intelligence scale of 500 million neurons and 250 billion synapses, ranking second in the world in terms of the number of neurons and synapses, and its energy efficiency is more than 10 times higher than that of existing computing systems. At the results release conference, the "Wentian" brain-like supercomputing team stated that it will continue to develop a new generation of brain-like computers, further innovate the brain-like computing chip architecture and software system framework, and create a brain-like computing platform that will lead future development.

Daily Economic News, Xinhua News Agency, Science and Technology Daily, Guangming Online

Daily Economic News