Cognitive computing is rapidly transforming big data analytics initiatives that began with reports and dashboards generated from nonstandard data into something more substantial. In fact, the cognitive computing market is expected to generate revenue of $13.7 billion by 2020, registering a CAGR (compound annual growth rate) of 33.1% during the forecast period of 2015 – 2020.
What’s cognitive computing, and how can you benefit from it?
Cognitive computing is a branch of artificial intelligence. It combines principles of science and engineering to produce “intelligent” machines that learn from the data they ingest in ways that hope to emulate the learning and thought processes of the human mind.
That being said, cognitive computing is not nearly as complex, sophisticated, and flexible as the human mind. Researchers discovered this almost 10 years ago when they tried to build a computer that could operate like the brain of a small mammal, and struggled with it.
Nevertheless, what cognitive computing can do well is exponentially multiple the powers of the human mind in solving elusive problems. For this, there is no better example than IBM Watson, and some of the transformative work it has done with big data in healthcare.
On an annul basis, approximately 5% of medical diagnoses made by physicians are in error. To attack this problem, and also to improve the process of diagnoses, Memorial Sloan Kettering, a leading cancer research center in New York, uses IBM Watson, which has analyzed over 605,000 pieces of medical evidence, two million pages of text, 25,000 training cases, and has had the assistance of more than 14,700 clinician hours fine-tuning its decision accuracy.
Here’s how IBM describes the process: A physician inputs symptoms and other related factors to the system, and Watson then identifies the key pieces of information and mines the patient’s data to find relevant facts about family history, current medications, and other existing conditions. Watson also reviews test results and forms hypotheses. What Watson comes up with in terms of diagnosis is not definitive, but the physician takes the Watson information and incorporates it into his/her own understanding of the condition and the patient he/she is treating.
This is a man-machine collaboration that takes advantage of the natural flexibilities, agility, and ability to innovate that characterize the human mind, and then combines this with a computer’s ability to process through reams of information in all forms that a human brain couldn’t hope to in any reasonable amount of time.
It is the tuning of this human-machine collaboration that enables the productive application of cognitive technology in companies. Here are three tips on how to do this to achieve optimum results.
1: Narrow your focus to a specific business case or application
The IBM Watson healthcare example was narrowly focused around providing more information and diagnosis support for physicians. Medical personnel using the application knew exactly what Watson’s role was in the diagnosis process, and how it would fit into their professional practice.
Likewise, businesses with other use cases should come to this narrow understanding of the “fit” or role of cognitive technology in their businesses. This begins with users understanding exactly how they will use cognitive technology and what benefits they should expect from it.
2: Make the application easy to use
User entries and exits into a cognitive computing work process should be simple and straightforward–whether these are by the click of a mouse or text-based input. Too much entry and exit “overhead” will discourage use, and diminished use of an expensive tool will not allow you to recoup your technology investment.
3: Set metrics and expectations and then measure results
If you are a transportation provider and your goals are to reduce maintenance and time of maintenance for equipment and infrastructure, a cognitive computing tool should be able to assess your assets, predict which assets need replacement or maintenance, and preventively recommend actions so your downtime and maintenance times are reduced.
At the beginning of cognitive technology implementation, you should set targets (e.g., reduce maintenance spend and time by 20%), calculate what your time and maintenance costs are prior to implementing cognitive technology, and then measure those costs after implementation to see if you are achieving the reductions that you projected. If you aren’t, you can further assess (or debug) your process to move toward expected performance levels.
The bottom line
At the end of the day, how you design the role of cognitive computing in your work processes and how you define the man-machine interface will be the determining success factors for your cognitive technology investment. That’s why getting the man-machine interface right is the most important thing you can do to improve the odds of capitalizing on cognitive computing.