Artificial intelligence (AI) works better when it learns like a child in its early stages.
This is what a study by researchers at the University of Aberystwyth says. Only when a machine starts small and figures things out on its own will it be ready for the difficult tasks that come with caring for the elderly, for example.
Robot learns by itself
According to study author Patricia Shaw, restrictions help you focus on specific problems. Only through practical experience would robots learn to cope with complex tasks and adapt to changing circumstances. To test how limitations affect AI learning, the team used a robot and turned off the function of its limbs. This means that the robot should only have limited control over its body, similar to a baby. The robot's camera was also manipulated and its image was made blurry so that the machine could see little, like a newborn.
The robot could decide for itself when it was time to reactivate the function of its limbs and camera. This way the robot could figure out how to move on its own. According to the researchers, this not only increased the speed at which the machine learned, but also the accuracy of the learning process.
[mk_ad]
Machines have to play
For the study, the researchers simulated the first ten months of a child's development. The machine learned the connection between his movements and his sensory impressions, displaying behavior similar to that of a child. For example, the robot often stared at its own hands during movements. By storing its experiences, the machine could learn from them on its own.
According to Shaw, AI is often programmed to do a specific activity. In order for the AI to learn these activities, they must first see thousands of images of specific actions, which requires a lot of time and effort. If instead we work with limitations at the beginning, a robot would get to know its own body much better. The next phase for Shaw would be playing. If a robot finds out how to interact with its environment through playful processes, it would later become even more adaptable.
Article image: Ryzhi / Shutterstock.com
Notes:
1) This content reflects the current state of affairs at the time of publication. The reproduction of individual images, screenshots, embeds or video sequences serves to discuss the topic. 2) Individual contributions were created through the use of machine assistance and were carefully checked by the Mimikama editorial team before publication. ( Reason )