Whether it’s learning how to build something or play a new sport, we all learn the same. Trial and error.
Researchers from UC Berkeley created new algorithms that bring this trial and error process to robots. A UC Berkeley press release describes the technique as a “major milestone in the field of artificial intelligence.”
This technique allows a robot to complete tasks such as putting a clothes hanger on a rack or assembling a toy plane without pre-programmed details about its surroundings.
Check out BRETT the Robot in action in the video below.
Alright BRETT, here’s some IKEA furniture. Good luck.
UC Berkeley Professor Pieter Abbeel touches on the positives of this technique. “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,” said Abbeel.
Abbeel and his fellow researchers will present this new technique at the International Conference on Robotics and Automation in Seattle on May 28.
Trevor Darrell, director of the Berkeley Vision and Learning Center, says robots must learn to ‘see’ and adapt to their surroundings before they can exist in our homes.
A traditional approach to helping robots navigate through our world is pre-programming a huge amount of possible scenarios the robot may encounter. The problem is; you can never account for every possible scenario.
Instead, researchers turned to a new sector of artificial intelligence called deep learning. It’s inspired by the neural circuitry in your brain.
A UC Berkeley press release does a good job of explaining what exactly deep learning is:
In the world of artificial intelligence, deep learning programs create “neural nets” in which layers of artificial neurons process overlapping raw sensory data, whether it be sound waves or image pixels. This helps the robot recognize patterns and categories among the data it is receiving. People who use Siri on their iPhones, Google’s speech-to-text program or Google Street View might already have benefited from the significant advances deep learning has provided in speech and vision recognition.
The researchers used a reward function to help BRETT learn how to complete tasks. Each movement is given a score based on how well the robot is doing the task. A higher score means BRETT is doing better.
Abbeel stresses the field of deep learning still has a long way to go, “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.”
I think it’s more a matter of when than if as far as AI goes. Notable tech companies such as Google are making a big push into the field of artificial intelligence.
Maybe in 10 years, I can give BRETT the Allen wrench and tell him to go to town.