The robot learned to decorate the dining table by looking at humans

The robot learned to decorate the dining table by looking at humans

The robot learned coding or trial and error tests have. Been used to teach robots. But teaching robots will soon be easier. MIT scientists have introduced a system called. Planning with Uncertain Specifications. In this way, robots can be taught complex tasks very easily.

In this system, the robot plans like a human to do something. And learns by watching humans. For example, decorating a dining table is a complex task for a robot.

For their work, the experts collected. A dataset of information on how to place eight pots on the table in different ways. The utensils included a mug, glass, spoon, fork, knife, dinner plate, small plate, and cup.

These pots can be placed on the table in many ways. For their work, the experts showed the robotic arm in various. Ways by arranging utensils on the table.

The experts then assigned the task of decorating. The table to the robotic arm. This robotic arm has learned from its observations. To decorate dishes on the table correctly many times.

Based on this system, the robot made. Only 6 mistakes in 20,000 attempts. Experts will soon teach the robot. Other household chores.