Only humans have the dexterity to accomplish many activities that animals, birds, and robots cannot do.
But things are changing with OpenAI bringing in the change. Elon Musk has a team of engineers who have taken up the challenge and come with a very close copy of the human dexterity.
The non-profit OpenAI of Elon Musk has come up with a robot that can manipulate objects with the same movement as the human hand, using Artificial Intelligence. The article used to teach the robot this dexterous movement was a child’s block.
Scientists at Open AI allowed the system to learn with different simulations to deal with the real world. A six-sided cube was provided to the robotic hand to manipulate it in various positions. By messing up the simulation, the robot was allowed to feel the cube and use it.
The AI system has been named Dactyl, and it is taught to grasp and change the cube with varying simulation gravity. Randomizations were used to teach the robot. It also meant that the computer required a high quantity of computing power.
Though movies show much dexterity in computer handling, it is still new in the real world. With research, OpenAI is coming with the required solution to make the computers more dexterous, while using their hands.
Another amazing fact that researches noted was the use of the thumb and the little finger by Dactyl. Dactyl preferred to use these two fingers more than the other fingers, unlike human beings who prefer to use the index finger or the middle finger. The fact that robots can discover and learn to adapt to its requirement has amazed the researchers at OpenAI.
A neural network was first developed by OpenAI, by which simulations were run at high speed. The same dataset was then used to grasp, slide and to manipulate the cube in many different ways with dexterity.