Cognitive Computing - The Next Era of Computing?
Cognitive computing is a subcategory of artificial intelligence that attempts to mimic the human brain, but it still has a long way to go.
Humans always tend to want more from life, and computing systems are not outside the purview of these expectations. From the time when computers were only able to tabulate data and then would do nothing beyond what they were programmed to do, we have strived toward creating computing systems that can find solutions to problems without human assistance. In a lot of ways, computing systems are now beginning to behave like human brains. This, known as cognitive computing, marks the beginning of a new era in computing.
Cognitive computing is a subset of artificial intelligence (AI), so cognitive computing has certain characteristics derived from AI, but there are still a lot of aspects in AI that have yet to be incorporated into cognitive computing. No doubt, this significant development is going to influence our lives like never before. However, the capability of computing systems to mimic human brains is also doubted by many. The human brain, as neuroscientists would agree, is highly complex and intelligent. In its present state, cognitive computing is able to mimic only an insignificant percentage of the human brain’s capabilities. (To learn more about computers trying to imitate the human brain, see Will Computers Be Able to Imitate the Human Brain?)
What Is Cognitive Computing?
Cognitive computing is the ability of computing systems to act like the human brain. Human brains can accept and store huge volumes of data in different forms such as text, visuals, sound, numbers and conversations. When required, the human brain can process the inputs and find solutions to situations and problems. Cognitive computing systems can perform similar tasks. It does not require data to be organized or compliant with a specific format when it is accepting data as input. After accepting information, it is capable of processing the information, organizing the data, finding patterns and making sense of such information. Based on what it has made out of the information it has received, it is capable of providing intelligent responses to questions. It does not stop receiving data or information and the information processing is continuous. A good example of a cognitive computing system is IBM’s Watson.
To get more clarity on cognitive computing, take the example below. It is a conversation between a health insurer’s cognitive system with a user who wants to buy health insurance but is not sure about which policy to buy.
Cognitive system (CS): Do you want to cover your family?
CS: How many members?
User: Three – myself, my wife and son.
CS: Give me the respective ages.
User: 40, 33 and 5.
CS: Do you have any preexisting health conditions?
User: Yes, I have asthma.
CS: In that case, you might have to pay extra premium.
User: What will the premium be?
CS: It will be $270. If you sign up for two years, you will get a 10 percent discount.
CS: I think Policy 1 and Policy 2 will be most suitable for you.
CS: Should I set up the automated premium collection system for you to pay the premium?
In the above example, the CS already has a lot of data on health insurance. Based on the data, it is advising the user on the most suitable health insurance policies.
Cognitive Computing, Machine Learning and AI
Both cognitive computing and machine learning can be considered subsets of AI. Obviously, there are certain commonalities across the three disciplines. Though machine learning performs tasks that are distinct, the differences between AI and cognitive computing are blurred. Many people are of the opinion that there is no basic difference between AI and cognitive computing because both enable computing systems to behave like human brains. But it is difficult for any one discipline to function totally independently of the other three. (For more on AI, see Thinking Machines: The Artificial Intelligence Debate.)
Machine learning is a type of data analysis that helps computers to find hidden insights from data it receives and build data models. Machine learning creates algorithms and fits them into computing systems so that whenever new data is received by the machine, the computing system could learn from the new data and build data models without any human intervention. Cognitive computing uses these data models to deeply understand new problems or situations and provides one or more solutions.
To summarize, to understand problems and provide solutions, cognitive computing relies on the data models, patterns and insights provided by machine learning.
Is Cognitive Computing the New Era of Computing?
Are computing systems today doing something significantly different from what computing systems did in the past? To enable computers to think and act like human brains, even partially, is a significant landmark – it is a machine, after all. When computers were invented, they were used to organize information in tabular form. After that, computers would be programmed to do specific tasks – without instruction, they would not be able to do anything. Now computers are able to learn from the inputs they receive without human intervention and provide solutions to problems.
To decide if cognitive computing has really ushered in a new era of computing, let's take a look at Airbus, one of the largest aircraft manufacturers in the world. With the help of cognitive computing and IoT, Airbus is able to gather important data about each aircraft that helps it to customize maintenance activities for each aircraft. Traditionally, aircraft would be brought in for maintenance at two-year intervals as per standards. But now, based on the data generated and processed by cognitive computing systems, Airbus can speedily fix issues and optimize performance. According to Laurent Martinez, head of business unit services at Airbus, data and cognitive computing are very important to the future of air travel.
Many of the newer models of the Airbus comprise 300 million parts which collect information about the aircraft during flight. While this information is used for better maintenance and upkeep of the aircraft, the cognitive computing systems within the aircraft also adapt to fix problems, if any, and optimize performance during the flight. This is also a huge cost-saving step. According to Martinez, "Overall, I'm deeply convinced that IoT, cognitive science and technologies will move aviation to the next level in terms of operations and in terms of passenger experience."
While the significance of the arrival of cognitive computing is being documented, a few questions need to be answered. There is still no clear distinction between AI and cognitive computing. If cognitive computing is indeed a subset of AI, what was the need to carve out a separate discipline of AI? While there are good reasons to praise the evolution of the computing systems, there is a lot of doubt as to whether a computing system can be as efficient as the human brain. A human brain is extremely complex and a supremely intelligent machine, and to equal it is no mean task. Until then, it would be appropriate to say that cognitive computing is partially capable of equaling the human brain.