You don't need a computer science degree to understand how enterprise machine learning is poised to change the way modern businesses do just about everything. That said, how is the average layman expected to separate the real from the exaggeration? According to Gartner, "machine learning" is nearly at the peak of the emerging technology "hype cycle."
Let's discuss what machine learning does and what it means for business with a couple oddly specific metaphors. And when we say "odd," we mean odd.
Machine Learning, the Baby Genius Going Through Growing Pains
Toddlers, born relatively tabula rasa, draw conclusions based on patterns they encounter on a daily basis. The ball is round. The dog goes "woof." The sky is blue. Together, all this information may tell babies they are, in fact, playing outside. What is "outside"? A place where they can play with balls, hear neighborhood dogs bark and see the sky.
Machine learning technology works as a hyper-accelerated version of this same pattern recognition paradigm. Like the average baby's developing cognitive faculties, enterprise tools equipped with machine learning algorithms attempt to glean actionable intelligence from the data they interact with.
As toddlers eventually grow into young adults, hungry for independence, so too will machine learning technology develop toward a state requiring little to no supervision. That may only come with time, but as Computerworld pointed out, many businesses have found notable success cultivating promising machine learning capabilities through collaborative "active learning" with trained employees. Forget Baby Einstein DVDs - as IT professionals assist machine learning models with recognizing algorithms and interpreting the data around them more thoroughly, their efficiency increases and they produce more accurate values.
Machine Learning, the Spoiler From a Movie With a Twist Ending You Never Saw
The Next Web recently ran an article featuring a machine learning app that can supposedly predict the chances each Game of Thrones character has of being killed off in the upcoming season.
That's certainly one method for explaining how machine learning works at a fundamental level, but let's highlight the limitations inherent in the system. Machine learning couldn't tell you how characters die, who kills them or when exactly they'll shuffle off the mortal coil. Granted, the app pulls data from the George R.R. Martin books the series is based on as well as online resources, but its primarily concerned with looking for characters' "attributes like their age, what House they belong to, whether they're married or how popular they are based on how many wiki pages link to them."
Machine learning technology, today and tomorrow, isn't really concerned with all the details in a consummate prediction. Rather, it forecasts several possibilities based on data values and past behaviors for users to interact with, then records their responses to get a little smarter.
And for businesses, that's more than enough for now. In digital workplaces utilizing machine learning to operate self-service ITSM solutions like help desks or service desks, connecting customers and users with a few different problem solving options is an important first step toward next-gen capabilities.