lswcap
A paper recently published by the Johns Hopkins University research team explores the possibility of learning through positive reinforcement under the title Good Robot. The key is to reward the robot for doing something right. If you do wrong, you will not be punished, and you will be rewarded for robots using a score method. If you do it right in the form of gamification of work, it gives you points.
The paper says that this method can reduce training time. Robots want to have high scores and learn the right actions quickly because they can get the best rewards. So far, if it took one month for the robot to do 100% accurate work, it could be reduced to two days.
The field of robot learning holds great potential for the industry. That’s one of the reasons why robots take so many different approaches to uncover the secrets so they can learn essentially from scratch. Of course, it is still only rudimentary tasks such as accumulating red eyes or playing video games, but it is expected that in the future, there will be a possibility to do more complex practically useful tasks. Related information can be found here .
Add comment