Over the decades, people have tried to copy a human mind, to use the software more and more sophisticated, creating a more sophisticated artificial intelligence system. But, despite all efforts, even the best artificial intelligence system is slightly larger than average intelligence cockroaches.
What is the problem? So far, all systems based on artificial intelligence software that works on the hardware has not changed. All changes to existing hardware to be done through software. But the development team from the University of Oslo (Oslo University) in Norway has taken a step which might form the basis of the next generation of hardware.
Professor Jim Torresen and Kyrre Glette built a developing hardware (Evolvable Hardware – EHW), which uses genetic algorithms. In other words, they created a computer “hardware” that can evolve.
System Torres and Gletta genetic means is that the hardware design, which can be most effective to accomplish the task. In the real world may need 20-30 thousand generations before the biological system will find a great configuration for solving a particular problem, and it can take 8-9 hundreds thousands of years. The system created by the Norwegian team spends the same amount of generation for a few seconds.
Work on EHW systems began in the late 90s. Torres Team began using evolutionary algorithms in 2004, when they made a robot chicken “Henriette”. Robot using evolutionary principles in its software in order to learn how to do some action, without attempting to understand the world and create a solution through the use of artificial intelligence.
Robot “Henriette” trying to apply random changes in their actions corresponding to the problem, and choose the best of them. So, after finding the best, it is able to solve their specific problems. Thanks to “Henriette” is to better understanding how evolution works. Similarly, the evolution of system hardware to work for finding configuration that will be most effective in solving problems.
Evolution can solve many problems that the programmer should generally provide capabilities to solve or address them. Let us assume the robot was sent to Mars and fell into the pit. Using evolutionary methods, the robot could learn how to climb out of the pit without human help.
Now the team wants to develop a robot to assist in the installation of oil pipes and other oilfield equipment at a depth of 2,000 meters. This depth makes it nearly impossible for a direct link with the robot. He should have 2-3 miles of wires extending behind it, or to transmit signals by means of the echo, which in turn will give a considerable delay between the command and its execution. Evolutionary robot could find a solution to problems encountered on site within a few seconds without operator intervention.