BEIJING, March 4 (Xinhua) -- Artificial intelligence technology is in the ascendant, and researchers have begun to explore organoid intelligence (OI) technology with more potential. One of the prospects, a biological computer powered by human brain cells, sounds like science fiction, but perhaps not far away, promises to reshape the future of modern computers and open up entirely new areas of research.
A woman experiences smart technology at the Consumer Electronics Show in Las Vegas on January 8. Xinhua News Agency (photo by Zeng Hui)
According to CNN reported on the 2nd, researchers at Johns Hopkins University in the United States conducted organoid intelligence technology research based on brain organoids cultured in the laboratory. They present their research ideas and progress in their latest academic journal, Frontiers of Science.
Brain-like structure with a large nib
Johns Hopkins University researchers have been trying to grow brain organoids since 2012. They extracted cells from human skin tissue and engineered them into embryonic stem cell-like structures for culturing brain-like organs. Brain organoids are the size of a pen tip and each contain about 50,000 cells, equivalent to the number of cells in the nervous system of a fruit fly. Thanks to neurons, brain-like organs have the potential to have basic functions such as learning and memory.
In a statement issued by Johns Hopkins University, principal researcher Thomas Hartung said that research with brain-like organs can avoid many ethical problems faced by using the human brain for research, and open up a new situation for the study of the working mechanism of the human brain.
Hartung and his colleagues conceived that using brain organoids as "hardware," they could potentially develop "biological computers" that are more energy-efficient than supercomputers. Hartung argues that while the contemporary technological revolution is driven by computer and artificial intelligence technologies, its development is "close to the ceiling."
Which is stronger, the human brain or the computer
Since the advent of computers, the question of which is more powerful than the human brain or the computer has arisen.
In terms of their ability to process data quickly and in large quantities, computers seem to far outperform the human brain. Take AlphaGo, for example, an AI Go software that "easily learned" 160,000 games, and it would take 175 years for a person to complete the same amount of training even if he studied 5 hours a day.
On May 23, 2017, the first showdown between Chinese player Ke Jie Jiudan and the Go artificial intelligence program "AlphaGo" was held in Wuzhen, Tongxiang City, Zhejiang Province, where Chinese player Ke Jie (left) was played. Photo by Xu Yu, a reporter from Xinhua News Agency
However, when it comes to complex logical problems, such as distinguishing between a cat and a dog, the human brain easily wins. In terms of energy efficiency, the human brain also performs better.
The supercomputer "Frontier" developed by the U.S. Department of Energy's Oak Ridge National Laboratory costs $600 million, weighs 3,629 kilograms, and each cabinet weighs the equivalent of two pickup trucks. Hartung said the computer's computing power did not exceed the computing power of a single human brain for the first time until last June, but it consumed 1 million times the energy required by the human brain. It can be seen that "modern computers still cannot be compared with the human brain."
The human brain is super capable of storing information, with an estimated 2500 trillion bytes of information. At the same time, the storage capacity of computers is close to the limit, and it is difficult for existing technology to add more transistors to tiny chips.
Research is just getting started
Researchers are therefore pinning their hopes on biocomputers and organoid intelligence technologies. Hartung defines organoid intelligence as "the reproduction of cognitive functions such as learning and sensory processing within a humanoid brain-like model grown in the laboratory."
He said that for the development of organoid intelligence technology, the brain organoids that can be cultivated are "too small", each only one 3 millionth the size of the human brain, and the memory storage is only 800 million bytes. In addition, organoid intelligence requires at least 10 million cells, compared to 50,000 cells per brain-organ.
On November 5, 2022, visitors experienced the interactive application of the metaverse in the artificial intelligence zone of the 5th CIIE Technology Equipment Exhibition Area. Photo by Xinhua News Agency reporter Jin Haoyuan
Researchers also have to figure out tricks: How do you input information into and read information from brain organoids? At present, they have developed a brain-computer interface device that resembles an "electroencephalogram cap" worn on a brain-like organ. It's a bendable shell covered with tiny electrodes that can both receive and transmit signals from and to brain-like organs.
The researchers also plan to apply bioengineering, machine learning capabilities, and other innovative techniques to organoid intelligence technology research.
Organoid research has gradually become large-scale in the past 20 years, and some researchers have cultivated small organoids in the laboratory that simulate the functions of human organs such as kidneys and lungs, mainly for research experiments, reducing human and animal experiments. Research on organoid intelligence technology is just beginning. Hartung estimates that it will still take decades for organoid intelligence technology to reach a level comparable to mouse brain power. (Yuan Yuan)