People and animals move quickly, often with remarkably little advance planning. In addition, we rely on highly developed and intelligent mechanisms that we use our senses to correct errors and to compensate. For example, when serving in tennis. We can train robots to act in this reliably? With the help of self-learning software, 14 robots and many test hours Google research has progress been made in hand-eye coordination of robots (see video). After 800,000 attempts knows how to distinguish the robot objects, and to improve the control of its own (robot) arm. This self-learning software allows robots able to move faster and more smoothly in the real world Source:. Google research Google with its Deep Mind artificial intelligence software not only working to physical actions optimize. In a battle with the world champion Go, a very complex Chinese board game, the Google AlphaGo program has won the first match. Despite some errors, the computer won at the end by choices that a man would never have claimed world champion Lee. That computers can calculate faster was known, but that they can be creative, gives a new dimension. Lee was surprised by the strong opening but also by unexpected choices Source:. The Verge – google alphago-go-deep mind-resulting 
Hand-eye coordination
No comments:
Post a Comment