Teaching Robots to Learn from Mistakes

Computer scientists train robots to learn from their mistakes.

13:59 10 November 2019

Computer scientists at the University of Leeds are using artificial intelligence techniques of automated planning and reinforcement to train robots learn from their mistakes. The goal is to enable the machine to assess unique circumstances presented in a task and find a solution – akin to a robot transferring skills and knowledge to a new problem.

 

 

 

Dr Matteo Leonetti, from the School of Computing said: “Artificial intelligence is good at enabling robots to reason – for example, we have seen robots involved in games of chess with grandmasters.

 

 

 

“But robots aren’t very good at what humans do very well: being highly mobile and dexterous. Those physical skills have been hardwired into the human brain, the result of evolution and the way we practise and practise and practise.

 

 

 

“And that is an idea that we are applying to the next generation of robots.”

 

 

 

Wissam Bejjani, the PhD researcher who wrote the research paper, added: “Our work is significant because it combines planning with reinforcement learning. A lot of research to try and develop this technology focuses on just one of those approaches.

 

 

 

“Our approach has been validated by results we have seen in the University’s robotics lab.

 

 

 

“With one problem, where the robot had to move a large apple, it first went to the left side of the apple to move away the clutter, before manipulating the apple.

 

 

 

“It did this without the clutter falling outside the boundary of the shelf.”