“Most existing robots are good at processing and performing the tasks they are pre-programmed to do, but they know little about the ‘real world’ where we humans live,” he said.
“So our project is an attempt to build a bridge between robots and that real world.”
The Self-Organizing Incremental Neural Network, or SOINN, is an algorithm that allows robots to use their previous knowledge to carry out tasks.
In a laboratory demonstration, when the robot was required to “serve water,” it began to divide the task into a series of previously learned skills: how to hold a cup, how to take a bottle, pour water from the bottle and place a glass down.
Thus, SOINN examines environment to gather data it needs in order to organize the information into a coherent set of instructions.
Hasegawa’s team’s goal is to try to merge these abilities in order to create a machine that searches the internet to learn how to do the required tasks.
“In the future, we believe it will be able to ask a computer in England how to brew a cup of tea and perform the task in Japan ,” he said.