Robot will know 'thinking'?

These robots can learn from experience and solve novel problems, just like humans. It sounds like a science fiction movie!

>>> Watch video: SOINN system helps robots serve water

But Japanese scientists have studied and made this possible. Dr. Osamu Hasegawa and his colleagues at the Tokyo Institute of Technology have developed a new generation of robots that can teach them to perform pre-programmed tasks and use items they have never seen before.

Picture 1 of Robot will know 'thinking'?
SOINN system helps robot analysis, reasoning (Photo: News)

The team has developed a system called ' SOINN ' - an algorithm that allows robots to use knowledge they already know to analyze and infer how to accomplish tasks. This system also helps the robot check the environment and collect the data needed to organize the information that it has received into a set of instructions.

" The current robots are able to handle and perform very well on pre-programmed missions, but they know very little about the real world where humans live ," said Dr. Osamu Hasegawa. AFP . " So our project is to build a bridge between the robot and the real world ."

For example, you ordered the robot to use the SOINN system: ' Let's serve water '. Though not programmed to serve water, robots will search the internet for similar tasks to meet the requirements step by step. SOINN systems can ask for help if tasks are delivered beyond the capabilities of the robot.

Picture 2 of Robot will know 'thinking'?
Image SOINN system for water supply

Just like humans, the SOINN system can filter out non-critical information that interferes with the robot during mission execution.

' The amount of information on the internet is huge, but robots can filter out unrelated results when looking for information, like human capabilities ,' 'said Dr Osamu Hasegawa.

Scientists hope the future SOINN system can be used in practice, such as driving traffic lights to avoid traffic congestion by analyzing traffic density data and reporting accidents. a specific time.