Artificial neural networks need to sleep peacefully to function as efficiently as humans.
By the end of this decade, experts believe that people will achieve a new momentum of technological progress, and the key to this progress will be artificial intelligence. Researchers around the world are working to accelerate the research and development of AI from its earliest stages to its most advanced form, where it will no longer need human assistance to operate and can actually do the same thing to humans. Make our lives easier. AI industry growth is expected to explode and its impact on business and society will begin to emerge. Given the current state of technology, there is still a lot of research and development that will enable AI to operate independently. Scientists are looking for new ways to help their technologies and applications retain more information to make them smarter than we are. Therefore, in trying to do the same, researchers have developed a new discovery that claims that artificial neural networks can retain information better if they sleep well!
Researchers at the University of California, San Diego School of Medicine claim that AI is just like humans. The human brain is very busy when it is asleep, repeating what we have learned throughout the day and can say the same thing to the artificial neural network. Maxim Bazhenov, Ph.D. And Professor of Medicine and Sleep Researcher at the University of California, Berkeley demonstrate how sleep develops rational memory, the ability to remember arbitrary or indirect relationships between human objects or events, and prevents us from forgetting old memories. . All in all, the brain works in a mysterious way, but no matter how it works, it can also help artificial neural networks in the same way to help machines maintain the same amount of knowledge and make them as productive and efficient as humans.
The catastrophic forgetfulness makes it difficult to retain knowledge in AI
Artificial neural networks generally use the architecture of the human brain to improve many technologies and systems, from basic science and medicine to finance and social media. In a certain way, they achieved great performance through computational speed, but these ANNs failed in one important aspect. They fail to learn continuously. It is a phenomenon called ‘catastrophic forgetfulness’ in which new information overlaps previous information or training. And if we consider the aspects of training, forgetting the catastrophe in AI can lead to major failures for researchers working on an important mission.
To eliminate this problem, scientists have used neural networks that can artificially mimic the human nervous system. Thus, instead of continuously entering information in a machine, it is transmitted as a separate event (surge) at a specific point in time. They found that these growing neural networks work better if they are paired with occasional lines that resemble machines that are sleeping like humans. Through the deployment of these methods, scientists have found that networks can be continuously studied and can avoid catastrophic forgetfulness on their own.
The evolution of AI is set to embark on the exponential growth of advanced computing, the deployment of cloud storage and memory, and the acceleration of smart device ecosystems. It is only a matter of time before people will be able to make the best use of AI. It seems that the elimination of catastrophic forgetfulness among machines is the first to achieve efficiency and progress.
Share this article.