A new study shows that systems controlled by the next generation of computing algorithms can produce better and more efficient machine learning products.
According to foreign media reports, researchers at Ohio State University found that digital twins (virtual copies) of electronic circuits showing chaotic behavior (chaotic behavior) were created by using machine learning tools, and the behavior patterns were successfully predicted and controlled using this information.
According to the journal Nature Communications, many everyday devices (such as thermostats and cruise control systems) use linear controllers that use simple rules to guide the system to the desired value.
For example, thermostats use such rules to determine the extent to which the space is heated or cooled according to the difference between the current temperature and the desired temperature.
However, because these algorithms are very simple, it is difficult for them to control systems that exhibit complex behavior (such as chaos).
Therefore, advanced equipment such as self-driving cars and aircraft usually rely on machine learning-based controllers, which use complex networks to learn the optimal control algorithms needed for efficient operation.
However, these algorithms have significant disadvantages, the most serious of which is that their implementation is very challenging and expensive.
Robert Kent, lead author of the study and a graduate student in physics at Ohio State University, says that achieving efficient digital twins could now have a wide-ranging impact on the way scientists develop future self-driving technologies.
, return to the first electric network home page >.
