Researchers develop new Algorithm to empower Robots to Learn like Humans New algorithms enable robots to learn motor tasks through trial and error, like humans. Researchers recently gave a demonstration for the technology by assigning a robot to complete various tasks, like placing a clothes hanger on a rack, screwing a cap on a water bottle, assembling a toy plane, etc. What was really impressive about the technique was the robot did not require pre-programmed details about surroundings.Pieter Abbeel said the new technique is certainly a new approach to empower a robot to facilitate learning. �The challenge of putting robots into real-life settings, like homes or offices, is that those environments are constantly changing. The robot must be able to perceive and adapt to its surroundings�, said co--researcher Trevor Darrell.The algorithm gives rise to a new branch of artificial intelligence, known as deep learning. The researchers chose Berkeley Robot for the Elimination of Tedious Tasks (BRETT) to take up a challenge of dealing with a relatively promising form of artificial intelligence called deep structured learning.The researchers have claimed that smaller amount of pre-programming is required when the algorithm is used in the robot. Also, it provides the capacity to work outside controlled environments like medical centers, factories or laboratories.A team led by Pieter Abbeel, an associate professor in the campus electrical engineering and computer sciences department, developed the new algorithm.Abbeel said the best thing about the technique is that it rids the need of reprogramming when the robot comes across something new. Basis of BRETT is the neural circuitry of the human brain which perceives and interacts with everything around it.Use of the algorithm is currently seen in voice recognition software, such as the iPhone's Siri.