Scientists are developing a new artificial intelligence technology that works closer to the human brain

Scientists are constantly trying to develop artificial intelligence and try to bridge the gap between it and human intelligence. In recent recent experiments, scientists have noticed that some AI programs have begun to work very closely with the human brain.

The study indicated that automatic networks work very close to the human brain.

A decade ago, scientists learned many of the most sophisticated AI systems use large stores of data to “train” an artificial neural network to correctly distinguish between things.

Such “supervised” training requires human sorting of data, which is very laborious, and neural networks often take shortcuts to learn to assemble things with minimal information, sometimes superficially.

For example, an artificial neural network (a group of computers connected together) can use the presence of grass to recognize an image of a cow, because cows are usually photographed in the fields.

A cross between animal intelligence and artificial intelligence

“Computers and artificial intelligence programs don’t really learn the subject, but they do a good job on the test,” said Alexei Efros, a computer scientist at the University of California, Berkeley.

Furthermore, for researchers interested in the intersection of animal and artificial intelligence, this “supervised learning” may be limited in what it can reveal about the workings of biological brains, since animals and humans do not use labeled data sets as their sole source of learning. does not use , but instead relies on Her experiences are based on her exploration of the environment on her own which leads her to gain a rich and powerful understanding of the world.

Today, some experts in computational neurology (the study of brain function in light of the information-processing properties of the structures that make up the nervous system) are beginning to investigate automated neural networks trained with little data that humans have categorized.

“Self-learning” algorithms for machines have proven very successful in learning human languages ​​(Getty Images)

Match animal brain functions

Machine “self-learning” algorithms have proven to be very successful in learning human languages, and recently they have succeeded in recognizing and distinguishing images.

In a recent study, computer models built to approximate mammalian visual and auditory systems and designed with self-supervised learning models for AI programs showed a closer match with brain function than their supervised learning counterparts.

For some neuroscientists, it appears that artificial networks are beginning to reveal some of the real methods that human and animal brains use to learn.

Neuroscientists developed simple computer models of a visual system using robotic neural networks when monkeys were shown the same images as opposed to artificial neural networks.

For example, the activity of real neurons and artificial neurons showed interesting similarities, and on one occasion scientists discovered models of communication between machines trying to detect sounds and smells.

Through repeated trial and error of artificial intelligence programs and connected neural robot networks, scientists are beginning to see a unique model for approaching the human method.

“I think there’s no question that 90% of what the brain does is self-supervised learning,” says Blake Richards, a computational neuroscientist at the AI ​​Cebic Institute.

And brains also learn on their own from their mistakes. Only a small part of our brain’s reactions come from an external source that tells us the answer is wrong.

Close results

Richards and his team created a self-supervision model for machines that help answer different questions. They trained an AI that combined two different neural networks: One, called ResNet, was designed to process images. A second network, known as the recurrent network, can focus on moving objects.

Richards’ team found that AI trained with RaceNet was good at recognizing objects but not at categorizing motion.

But when they split the communication network into two parts, this led to the creation of two paths (without changing the total number of neurons), the AI ​​developed a section for recognizing static objects and another for moving objects , which finally allowed it to classify the scenes presented to it, which according to scientists is the method by which our human brains work.

To further test the AI, the research team showed the artificial neural network and a group of mice a number of videos. It is worth noting that the brain of mice has areas of the brain that are specialized in static images and others that are characterized by movement.

Finally, scientists have confirmed that the human or animal brain is full of so-called feedback connections, while current models of AI have very few, if any, of these connections, raising a critical question about how advanced AI programs are. main distinguishing factors of the human brain.

Leave a Comment