What The Go Computer Battle Can Teach Us About Big Data

Original article is posted at Forbes : LINK

People love a good man-versus-machine story. From Deep Blue vs. Kasparov in chess, to Watson vs. Ken Jennings in Jeopardy, there is tremendous fascination in pitting humanity’s best and brightest against ever more powerful technological adversaries.

Most recently, Google grabbed headlines when its AlphaGo program defeated 9 dan-ranked professional Go player Lee Sedol. Much attention has rightly been given to the remarkable accomplishment of the Google team, whose use of deep neural networks combined with an advanced tree search algorithm was brilliant and a mark in the history of computer science.

As someone who is familiar both with the culture of Go and deeply immersed in big data, another technology front where achievement grows by leaps and bounds, I was impressed, but also deeply struck by parallels between the two worlds and the negative implications of some of those similarities. In particular, the AlphaGo project may just be the perfect roadmap for how not to approach Big Data in a business environment.


In countries like South Korea, China and Japan, Go is more than a game. Like soccer parents who drive their children down a path to be the next Hope Solo or David Beckham, the road to Go greatness is an incredibly competitive journey that for some begins as early as age 3. Parents pour resources into training—coaches, exclusive schools, and competitive circuits—in the hopes that their child, by age 17, will emerge as one of the rare professional Go players. The number of people achieving professional status each year is in the low single digits, but for those who succeed, tremendous prestige is within reach.

But what of the child who doesn’t turn pro? After 14 years of nothing but playing Go, many are socially inexperienced and have little in the way of other skills. Of course, they are smart and they’ll likely go to college, but so much of their experience and ability is predicated on what suddenly becomes a rather esoteric skill.

To achieve its Go victory, Google too made a massive, tightly focused, singular investment in terms of money and top-level PhD brainpower. While Google harvested some marketing success and gained valuable technology developments, there was no assurance of success. While Google may be able to write off millions, most other companies can ill afford such massive opportunity cost with little assurance of a payoff.

While the handful of companies most frequently cited as the wizards of big data—Facebook, Amazon, LinkedIn, and Google—have achieved remarkable results, their “moonshot” approach is not one that most people or companies can or should even try to duplicate. What if it doesn’t work?

Rather than an exclusive cadre of data scientists locked away in a clean room working to capture a one in a million insight, it is more effective to train as many people as possible to experiment for themselves. What we really need is to empower people with a wide range of experience in all facets of the business, the industry, and the marketplace to ask creative questions of big data. This democratization of big data using increasingly accessible tools is a smart investment in the entire business that breeds a culture of far-reaching inquiry that can make more flexible and innovative use of big data than any three PhDs in a room full of servers can achieve on their own.

Like the Go genius in training, Big Data that’s locked away and kept on a single track is an all-or-nothing approach. The outcome is binary: success or failure. The way to reap victories from big data is to create a culture where everyone can experiment and become big data literate, with each person bringing his or her unique perspective and creativity to bear on solving business problems.

Views: 472


You need to be a member of Data Science Central to add comments!

Join Data Science Central

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service