Position: Home > News > capital
Robots are also curious and can set their own learning goals
Core hints :In the past few years, machine learning technology has made great breakthroughs. After Alpha Go's victory over human players, attention to such technology has never been higher. At the same time, the autonomous driving technology has also realized th

Release date:2019-11-13

Browse number:45

 In the past few years, machine learning technology has made great breakthroughs. After Alpha Go's victory over human players, attention to such technology has never been higher. At the same time, the autonomous driving technology has also realized the rapid development, the driver's hands have been liberated to a certain extent. Even so, there are many tasks that robots can't perform due to a lack of daily basics, such as cleaning a room.
 
To address this problem, a research team at the Italian national research council's institute of cognitive science and technology (istc-cnr) has developed a goal-based open automatic robot learning program (GOAL), which aims to equip a GOAL robot with childlike curiosity and then drive it to teach itself. It is reported that this project will be a major breakthrough in the field of AI.
Robots are also curious and can set their own learning goals
 
The project's main researcher Gianluca Baldassarre said in an interview with The Express, he will call outfits, participate in The innovation project of research and development: "our common goal is to let The robot can possess a curiosity like children aged one to two, and then in The process of self-study, will they learn new technology to good use." In other words, robots can set their own learning goals, and then they will develop curiosity, which will drive them to learn knowledge.
 
Where, then, will Baldassarre's reference to "all these people" be responsible for research and development?
 
First, the paris-based team will focus on how children set their own learning goals and how people motivate themselves to teach themselves. They will then expand on a previous parent project called im-clever, which studies intrinsic motivation in human learning (IM) and USES the open-source iCub robot as hardware.
 
 
 
In addition, two other research and development teams in Germany will continue to develop the hardware needed for the project. The Frankfurt team will develop a new vision system and "brain", while a roboticist from darmstadt will work on the prototype.
 
In general, Baldassarre hopes that the development of the Goal robot will enable the robot to adapt and learn autonomically in any environment, thus solving many of the problems that currently beset roboticians. In a recent Robohub post with Vieri Giuliano Santucci, he explains:
 
8.0
 
"If curiosity and intrinsic motivation are the basis for different skills and adaptability in humans, having algorithms similar to intrinsic learning in a robot would allow a 'learning engine' to drive the robot well in an open, automatic learning process. This process does not require any programming or training on the part of the scientist."
 
The robot will eventually carry out tests to organize cabinets. It may seem boring at first, but it requires the robot to be able to see, grasp, and understand how objects stand, and how to clean and organize more efficiently in the midst of chaos.
 
So far, the project has raised nearly $4 million, and the test results will be shown in 2020.
 

Friendship link    | International Artificial intelligence

News Exhibition Video Product Enterprise Picture Technology
Liaoning ICP No. 234234234 -1 | Liaoning No. 1102342342030919 public security