Scientists will connect brain cells and artificial intelligence: it could be a revolution
In Australia, a group of researchers has received funding for a project in which they plan to combine the human brain with artificial intelligence. Earlier, the same researchers had already taught a test-tube-grown brain to play ping-pong on a computer.
According to a press release from Australia's Monash University, which will be in charge of the project, a grant of more than $400,000 has been allocated from the country's budget. The research will be conducted in conjunction with Melbourne-based startup Cortical Labs, which previously "made" about 800,000 brain cells in a petri dish to play ping-pong on a computer.
The idea of combining biology with artificial intelligence, scientists believe, could open new frontiers of machine learning. This could later be used in self-driving cars, autonomous drones, delivery robots and more.
"This new technological capability could surpass the performance of existing silicon chip-based hardware in the future," said team leader and Monash University associate professor Adil Razi.
He believes the research findings will have a significant impact on different fields such as planning, robotics, advanced automation, brain-machine interfaces and drug development.
The main advantage of building artificial intelligence based on the human brain is that it is likely to be capable of "lifelong learning" like human brain cells. This will allow AI to learn new skills without losing old ones, and to combine existing knowledge with new tasks.
Modern AI cannot do this and suffers from "catastrophic forgetting". In contrast, the human brain is capable of lifelong, continuous learning.
Razi and his colleagues aim to grow brain cells in a lab dish called DishBrain to investigate the process of "continuous lifelong learning."
The researchers note that this is a very ambitious project that is likely to take a lot of time, so don't get your hopes up for quick results.
"We will use this grant to develop better artificial intelligence machines that replicate the learnability of these biological neural networks. This will help us scale hardware and methods to the point where they become a viable replacement for in silico computing (on silicon chips. - Ed.)," Razi emphasized.
Earlier, researchers at biotech startup Cortical Labs created a "mini-brain" consisting of about 800,000 living human brain cells in a petri dish. The cells are arranged on top of an array of microelectrodes that analyze neural activity.
"We think it's fair to call them a cyborg brain," Cortical Labs chief scientist and scientific director of the project Brett Kagan told us.
For their "test subject," the scientists created a simplified computerized version of ping pong with no opponent.
Then they found that the mini-brain was able to play this kind of tennis for one very quickly and its learning abilities were better than those shown by the AI.
"The amazing thing is how fast they learn - in five minutes in real time. It's really an amazing thing what biology can do," Kagan asserted.
Earlier, OBOZREVATEL also reported that Ilon Musk's Neuralink company received permission to implant a chip in the human brain. The decision was made despite the fact that preliminary tests killed more than 1,500 animals.