Scientists from Cornell University in New York have created a Robo Brain that has the capacity to learn and, in turn, teach other robots to better serve humans.

The Robo Brain is defined as a large-scale computational system that does all its learning from available internet resources.

Like Us on Facebook

Currently, it's downloading and learning from a billion images, 120,000 YouTube videos and millions of how-to documents and appliance manuals.

These pieces of information are being translated into formats that robots can understand.

The way it works is that it functions like a hive mind on the internet.

Robots and artificial intelligence -- like Siri -- will be able to draw information from Robo Brain to understand how certain things work, how humans talk and how to handle things.

Lead Researcher Ashutosh Saxena says that since our smartphones and laptops can access the internet, we immediately have access to all the information we nwws.

In the same way, robots will be able to access the Robo Brain if it encounters things it's not familiar with. It can draw information from images, videos and whatever is stored in its memory banks.

Thus, robots will know what a chair looks like and not to step in front of a human watching TV

This system employs a structure called "deep learning." This process that stores information in many different levels of abstraction.

Different pieces of information are stored in separate nodes, each one being assigned a probability as to how much the information can vary but still be correct.

This allows robots to search for knowledge; connect chains of information and see which of them correctly applies to the situation at hand.

Saxena adds that the goal of Robo Brain is not to make a profit, but rather, to make the world a better place.

TagsRobo Brain, Ashutosh Saxena