Russian Scientists Claim Artificial Intelligence Breakthrough

Artificial Intelligence  

Russian scientists claim they are one step closer to creating full artificial intelligence. A physical model of a brain has been designed, with the ability to educate itself.

Researchers in Russia claim to have made a major step towards the goal of artificial general intelligence (AGI). They have created an "artificial medium of natural intelligence - physical model, capable of self-learning."

The international team of scientists at a laboratory based in Tomsk State University in western Siberia have created a device that could be an artificial carrier of a natural mind, able to learn and react to the environment, according to a press release.

The system was built based on mathematical and computer models of the human brain states the head of the effort, Professor Vladimir Syryamkin. "After that it was designed radio-electronic device comprising perceptrons. It is capable of handling various information (video, audio, etc.) Now we are working to establish the basic system robotic system, which is an intelligent control center.

"In the final version, we want to achieve maximum proximity of artificial intelligence to the biological model of the brain."

"In the end, an artificial brain should be analogous to the biological model," the main developer of the system, Shumilov Vladimir said. "We have a tremendous amount of work, but a very important step has been taken - we were able to reveal the secret of brain neural network. In our physical model, as in the human brain, the formation of new neural connections and damping existing. In humans, it is the process of forgetting."

This physical model is capable of self-learning and life experience.

Related articles DARPA Looks To New Form Of Computation That Mimics The Human Brain  Robin Hanson Envisions a Future Robot Society  Stuart Armstrong and Nick Hawes Discuss the Future of Artificial Intelligence

This mechanism is both simple and complex. The artificial medium of natural intelligence takes external stimuli such as light, sound, etc. Through trial and error, it tries to find a solution that helps to avoid the impact of the stimulus.

According to the researchers, when exposed to a bright light source the system will first try to look away, if it does not sense that it has moved enough, it will continue to move away from the light. As long as the artificial brain does not find the right solution, its neurons (perceptrons) will be in an excited state. When the artificial intelligence that decision will, it will remember it and will be used in similar situations.

"In the final version, we want to achieve maximum proximity of artificial intelligence to the biological model of the brain," says Syryamkin . "Therefore, in a short time in our project will join experts from other fields like biology and psychology. After all, artificial intelligence should not only be able to think for themselves, but also to distinguish between what is good and what is bad from the point of view of ethics."

According to the researchers, the possibilities created by the use of artificial intelligence are almost limitless.

SOURCE  Tomsk State University

By 33rd Square

Embed