Cognitive computing is primarily a marketing term indicating a computing service that is able to understand, reason and learn from the data it is supplied with. In essence it is the application of machine learning and artificial intelligence to data processing. IBM is the flag bearer for cognitive computing. Presumably it wanted a term that differentiated its Watson cloud based service from the ocean of other such services. IBM has its own definition of cognitive computing, shown below:
“Cognitive computing refers to next-generation information systems that understand, reason, learn, and interact. These systems do this by continually building knowledge and learning, understanding natural language, and reasoning and interacting more naturally with human beings than traditional programmable systems.”
So, if you are now floundering in a sea of marketing-speak, let’s get specific. For the time being we will dispense with the term cognitive computing and simply talk about AI. There are dozens of very specific techniques and methods that come under the AI banner, from processing natural language through to finding the best route between Chicago and New York. The holy grail of AI is called Artificial General Intelligence (AGI). This would be able to handle most problems where intelligence is needed using a single method. It’s decades off, and there isn’t even a widely accepted definition of the term. Real cognitive computing needs AGI, otherwise it’s just a fancy umbrella term for traditional AI.
Lets strip off the marketing gloss a little more. AI methods generally employ brute force to process data and deliver some form of intelligent functionality. There is nothing particularly intelligent about it – other than the intelligence of the people who created the various algorithms. Things are moving forward, and deep learning is proving to be remarkably effective in image recognition – but it is still brute force.
The term cognitive computing may well become broadly used – that’s the power of billion dollar marketing budgets. And presumably suppliers like IBM feel they need a term that smells as little as possible of technology – they are selling to business managers after all. But for those who want to know what they are buying, the reality for now is that cognitive computing is AI and machine learning, and still requires technical people who know what they are doing. The dream of a voice activated interface that will perform any task you ask of it, is just that- a dream.