The speed of change in the digital economy

Cognitive systems: a new era of computing

Cognitive computing refers to the development of computer systems modeled after the human brain. Cognitive computing integrates technology and biology in an attempt to re-engineer the brain, one of the most effective computers on Earth. Cognitive computing integrates the idea of a neural network, a series of events and experiences which the computer organizes to make decisions.

Inside Google’s secretive X laboratory, known for inventing self-driving cars and augmented reality glasses, a small group of researchers began working several years ago on a simulation of the human brain. There Google scientists created one of the largest neural networks for machine learning by connecting 16,000 computer processors, which they turned loose on the internet to learn on its own. Presented with 10 million digital images found in YouTube videos, what did Google’s brain do? What millions of humans do with YouTube: looked for cats.

The pretty amazing takeaway here is that this 16,000-processor neural network, spread out over 1,000 linked computers, was not told to look for any one thing, but instead discovered that a pattern revolved around cat pictures on its own. The network itself does not know what a cat is. What it does realize, however, is that there is something that it can recognize as being the same thing, and if we gave it the word, it would very well refer to it as “cat.” The system had a digital thought.

The Google research provides new evidence that existing machine learning algorithms improve greatly as the machines are given access to large pools of data. The size of the network is important, too, and the human brain is a million times larger in terms of the number of neurons and synapses than Google X’s simulated mind, according to the researchers.

Google researchers are not alone in exploiting those techniques, which are referred to as “deep learning” models. In 2011, Microsoft scientists presented research showing that the techniques could be applied equally well to build computer systems to understand human speech.

Deep learning methods are based on two ideas:

• Learning multiple levels of representation in order to model complex relationships among data, and 
• Learning representations from unlabeled data.

Whereas in today’s programmable era, computers essentially process a series of “if then what” equations, cognitive systems learn, adapt, and ultimately hypothesize and suggest answers. Cognitive systems are designed for statistical analytics. Watson, the Jeopardy-winning system from IBM, is an early example. When Watson answers a question it analyzes uncertain data, and develops a statistical ranking and a level of confidence in its answers. 

Potential applications include improvements to image search, speech recognition and machine language translation.