Last year, scientists at Monash University created "DishBrain" - a semi-biological computer chip with about 800,000 laboratory-grown human and mouse brain cells on its electrodes. According to New Atlas, at the time, the chip demonstrated a certain sentience and learned to play Pong in five minutes.
DishBrain's array of microelectrodes was able to read the activity of brain cells and stimulate them with electrical signals. That said, the research team created a version of Pong in which brain cells received an electrical stimulus in motion to represent the side of the "screen" the ball was on and the distance it was from the racket. This allowed the brain cells to act on the racket, moving it left and right.
Then they created a very basic reward system, using the fact that small groups of brain cells tend to try to minimize the unpredictability of their environment. Thus, if the racket hit the ball, the cells would receive a pleasant and predictable stimulus. If he didn't get it right, the cells would receive four seconds of totally unpredictable stimulation.
This was the first time that brain cells grown in the laboratory were used for this purpose, being given not only a way to feel the world, but also to act on it.
The results were considered so impressive that the research, carried out in partnership with Cortical Labs of Melbourne, has now attracted a grant of US$407,000 from Australia's National Intelligence and Security Discovery Research Grants program.
Chip that fuses biological computing and AI
The results of this research will have significant implications in a number of domains such as, but not limited to, planning, robotics, advanced automation, brain-machine interfaces and drug discovery, giving Australia a significant strategic advantage.
Associate professor and project leader, Adeel Razi, said.
The same source added that these programmable chips, which merge biological computing with Artificial Intelligence [AI], "in the future, may surpass the performance of existing hardware, purely based on silicon".
In the professor's opinion, this type of advanced learning capabilities could give autonomous vehicles, drones and robots "a new type of machine intelligence, capable of learning throughout its life".
We will use this grant to develop better AI machines that replicate the learning ability of these biological neural networks. This will help us to increase the capabilities of the hardware and methods to the point where they become a viable replacement for silicon computing.
0 Comentários