The Rise of Machine Learning

Big-Data-2-compressorJim Scapa has been chairman and CEO of Altair Engineering for 12 years, but he’s having difficulty getting others at the Troy, Michigan-based design and development engineering company to grasp the importance of machine learning.

“I’ve suggested using it for a couple of things here recently but got persuaded otherwise,” he says. “But I still think I’m right, and I’m really persistent.”

What Scapa recognizes is that machine learning “leverages the fact that there’ s a huge amount of computing available to try to solve problems, predict the future and prescribe solutions.”

Indeed, Big Data is revolutionizing business, much as the Internet began doing a quarter-century ago, and machine learning is emerging as one of the most important tools for utilizing it. Just as with the web before it, CEOs like Scapa realize they must use machine learning to gain a competitive advantage for their companies—or they’ll most certainly be seeing their rivals do it instead.

Machine learning is a branch of artificial intelligence that enables machines to learn on their own, without much human supervision, drinking deeply from the well of Big Data. Computers
essentially write and follow their own programs based on the statistical relationships they discover in unstructured data—and are roiling industries ranging from credit cards to automobiles
in the process.

“It’s speed and the ability to learn from data that gives machine learning the power to provide tremendous insights in ways that humans could never do on their own or with basic business-intelligence tools,” says Mike Tuchen, CEO of Talend, a Los Altos, California-based big-data integration firm.

Or, as Altair’s Scapa put it, machine learning “can use algorithms to mine historical data for outcomes that are different than with traditional simulation.” At financial companies, for instance, machine learning can assess insider-trading activities and identify potential fraudulent activities that could trip a regulatory investigation. At utilities, machine learning can work with digital smart meters to identify sources and patterns of energy consumption in a house so that the company can “generate revenue around that data,” says Josh Sutton, the global artificial-intelligence lead for Sapient consultants.

Consider how machine learning is improving credit-card authorization by leaps and bounds. “The system learns by itself and actually keeps evolving algorithms by itself, as authentication keeps automatically refining on its own,” says Raja Rajamannar, chief marketing officer for New York-based MasterCard. “So the accuracy of what we’re doing increases, while the false positives and false negatives keep declining.”

Or look at what True Fit is doing with machine learning. The Woburn, Massachusetts-based startup amalgamates all sorts of data about an individual consumer, including purchase patterns,
to curate online clothing selections just for them, and machine learning makes TrueFit more helpful as it learns more about the customer.

Big-Data-3-compressor“It’s kind of like having a Pandora profile for music,” says CEO Bill Adler. “The data tells us your body resembles this, this is what you wear, here are the colors you like, so we can put together special recommendations and a really incredible digital experience for the consumer.”

Automatic Data Processing (ADP) has figured out how to use machine learning to advise clients which employee time cards need to be reviewed and which don’t, based on previous diligence patterns of the employees—sort of like the “pre-check” feature for travelers deemed safe that the federal Transportation Security Agency initiated at airports.

Dale Buss
Dale Buss is a long-time contributor to Chief Executive, Forbes, The Wall Street Journal and other top-flight business publications. He lives in Michigan.