Five Myths About Machine Learning You Need To Know Today

Original article is published at Forbes: Link

Ask most people outside academia or Silicon Valley what comes to mind when they hear the term “machine learning” and you’re likely to get a response that involves a movie like “The Matrix” or “Ex Machina.” You’re less likely to hear how it’s a great tool for fraud detection or supply chain optimization, and that’s too bad. Machine learning has a tremendous range of business applications, from optimizing data centers to predicting fine wine price changes to retail market basket analysis. With that in mind, I hope to cut through the science fiction clutter and misconceptions so you can consider how machine learning relates to your business.

mythMyth 1: Machine learning is only for PhDs

Many have heard about Andrew Ng’s popular graduate level machine learning course at Stanford, now available on Coursera. You won’t need to take that course to benefit from machine learning (although admittedly it’s a great course). Learning to write machine learning algorithms is quite different from learning to use them. After all, you don’t need to know how to program an app to use your iPhone. As with many technologies, the best platforms abstract away the obscure to present business users with applications that require minimal training. If you know your use case and basic concepts of machine learning, you are ready to go. The technical expertise to fine tune which algorithms make the most sense for a particular use case is left to data scientists. Users don’t need to know the math, just their business domain.

Myth 2: Machines will take over

All those movies where the robots understand themselves, learn through experience, and seek to dominate humanity? That is not machine learning! That is artificial intelligence and the distinction is important. Machine learning is a tool you can choose to accept and put to use. Just like you have the choice to use Facebook or not, there is a decision to be made and choices that determine how much it influences your life. My colleague John Thuma got it right when he described machine learning as the next penicillin. Machine learning is something that extends your capabilities, not the machine’s.

Myth 3: Machine learning is just a marketing buzzword

On the contrary, machine learning has a long history. Legendary computer scientist Arthur Samuel defined machine learning as early as 1959 in his seminal work on computer checkers as a “field of study that gives computers the ability to learn without be...” Some suggest the term is far older, dating back to the 17th century. Machine learning is used today in mainstream products including Google search, Netflix recommendation algorithms and Nest thermostats. The simple truth is machine learning is not hype. It is real, practical, and making money.

Myth 4: Machine learning has nothing to do with my business

You might think, I’m not into robotics, so machine learning doesn’t apply to me. Machine learning has applications in a wide variety of general use cases. Machine learning is everywhere and has applications any time we are trying to make predictions based on diverse signals from multiple sources, whether they are sensors attached to an oil well thousands of feet below the surface or the click of a button on a mobile shopping app. From predictive maintenance to demand forecasting to preventing fraud, machine learning can be used to help solve a wide variety of business problems.


Myth 5: Machine learning can solve all my problems

On the other hand, machine learning is not a silver bullet, capable of solving extremely difficult, unsolved problems. For that you need complex algorithms. At the other end of the spectrum, if you know which dimensions to use, you don’t need machine learning. But if you don’t know which dimensions to use and there are endless possibilities, that’s a great use case for machine learning. You may get weary of testing new possibilities, but machines don’t.

Consider predicting buying behavior. Such behavior might be influenced by an almost unlimited number of factors and that list of factors might be different for different customers or different for the same customer under different circumstances. Machine learning offers the ability to tirelessly iterate, trying out additional dimensions until we find which factors really are predictive for a given customer.

Machine learning is not science fiction. It is not even out of reach. With a little understanding and the right expertise, machine learning is ready to be harnessed by curious, competitive businesses of all descriptions.

Views: 3939

Tags: learning, machine


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

Join Data Science Central

Comment by Serge Pvoti on May 1, 2017 at 11:56am

As the value of machine learning grows, we are faced with vast opportunities that will change the world, but we need to guard against "cargo cult". Through  poor understandings of statistics/machine learning, we risk creating an industry defined by unreliable results.

Comment by Nitin Sareen on June 30, 2016 at 10:11am
Leaving the fine tuning to data scientists and letting non-trained users amok with tools is not the right approach. It is important for a user to understand what technique needs to be used.
The "I-don't-need-to-know-the-algorithm-but-I-can-do-machine-learning" approach will create a workforce that is not ready and will cause the downfall of machine learning.
Comment by Sione Palu on June 27, 2016 at 12:45am

Machine learning is applied mathematics & that's it.  The idea that a machine can learn in a  humanlike manner is nonsense.

© 2021   TechTarget, Inc.   Powered by

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