Introduction: What is Cognitive Computing?
The possibility of something other than programmable systems was predicted as early as 1960 by JCR Licklider, a pioneer in the computing field. In his influential paper “Man-Computer Symbiosis” Licklider envisaged that cognitive computing would become a future necessity, and today much of modern computing is based on his research.
Cognitive computing refers to mimicking the workings of the human brain in a computerised model. It comprises self-learning systems that can make sense of data using artificial intelligence and algorithms, enabling it to even think and learn. Cognitive-based systems like IBM’s Watson have the ability to understand natural language, rationalise and relate to human beings.
Instead of being programmed to await an answer or action like traditional computers, a cognitive computing system develops its decision-making and logic capabilities with each interaction. With repeated interactions, a cognitive computer system will learn who you are, your likes and dislikes and can adapt its output of information to be beneficial to you.
Capabilities of cognitive computing systems
A recent study from the IBM Institute for Business Value, “Your Cognitive Future”, pinpoints three types of capabilities that cognitive systems can offer users:
- Engagement – These systems are like a tireless assistant who can merge unclear data and present information in a timely, practical manner.
- Decision – Cognitive computing systems have the ability to ‘advise’ a human to make a final decision by suggesting options based on evidence.
- Discovery – These systems can find insights that could possibly not be discovered by even the most skilful human beings. Progress in this area has been made in fields such as medical research where there is a vast amount of information available.
Cognitive computing problems
Pushing the evolution of cognitive-based systems, particularly when it comes to algorithms, is proving to be a real challenge for companies like IBM. However, the difficulties have less to do with technology limits, but rather with next-generation development challenges that are rapidly developing.
Says Jerome Pesenti, vice president of the Watson team at IBM, “When it comes to neural networks, we don’t entirely know how they work…And what’s amazing is that we’re starting to build systems we can’t fully understand. The math and the behaviour are becoming very complex and my suspicion is that as we create these networks that are ever larger and keep throwing computing power to it, it will create some interesting methodological problems.”