The selection process involved in this study is similar to the natural selection, but the selection criteria are larger and faster than natural selection. This allows the artificial brain to select most beneficial features and evolve. Leading the research is Dr. Chris Adami, a computational biologist at Michigan State University, who uses genetic algorithms to mathematically model a large group of machines "brains" that perform a task. For example, the task could be to find the exit of a labyrinth. The results of the modeling experiments show that the most complete machine "brains" with the most complete tasks produce the largest number of virtual "offsprings", a result that means the most intelligent robots can "reproduce." Researchers have used this genetic algorithm to select the "brains" of the machine for thousands of generations, sometimes hundreds of thousands of generations, and download the "live" "brains" to the robots, and then let the robots perform the reality. Various tasks in the world. Robot brain One of the many tasks performed by these robots is the most complicated. This task requires multiple robots to figure out and remember the order in which they walked out of a room. The scientist then ordered the robots to re-enter the room in the order described above, or in the reverse order. "This task is complicated because these robots must be able to identify each other's identity," Dr. Adami said. After running this genetic algorithm for selection, these robots seem to solve this problem, and they have learned to use certain actions to tell other robots their identity. Dr. Adami believes that it is the best way to generate self-aware artificial intelligence by letting the robot "brains" evolve in a complex world and forcing them to interact. Es Control Cabinet,Customization Es Control Cabinet,Electrical Meter Control Cabinet,Outdoor Es Control Cabinet Huaian Qiangsheng Cabinet Co., Ltd. , https://www.qscontrolcabinet.com