UC Berkeley Researchers Make BRETT The Robot Learn Like a Child

Researchers at UC Berkeley have taught a robot to learn by erring, also known as 'trial and error'.

The Berkeley Robot for Elimination of Tedious Tasks (BRETT) could master unfamiliar tasks like placing blocks into matching openings, stacking Lego blocks and unscrewing a bottle in 10 minutes with some assistance. It took as long as three hours without help. Researchers say increased processing power can shorten learning time.

"With more data, you can start learning more complex things. We still have a long way to go before our robots can learn to clean a house or sort laundry, but our initial results indicate that these kinds of deep learning techniques can have a transformative effect in terms of enabling robots to learn complex tasks entirely from scratch. In the next five to 10 years, we may see significant advances in robot learning capabilities through this line of work," said Pieter Abbeel a professor at the university.

Abbeel's team used a technique called Deep Learning, which has been famously used in Apple's Siri and Google's Speech to Text programs. While deep learning is easily applied to images and speech, researchers say getting a robot to use it in a 3 D environment is complicated.

Researchers relied on a layer of artificial neurons built in the robot to use raw data from surroundings. BRETT then looks for patterns in the data to help with tasks. Algorithms guiding the robot associated a score with the way it performed a task. Movements which helped score high provided feedback to the neural networks and helped in mastering the task. The approach is similar to mimicking humans' trial and error learning approach.

"The key is that when a robot is faced with something new, we won't have to reprogram it. The exact same software, which encodes how the robot can learn, was used to allow the robot to learn all the different tasks we gave it," Abbeel said.

The team will present the technique at the International Conference on Robotics and Automation in Seattle on May 28.