There are a number of concepts that you might, or might not, have heard of in recent months. Big data, business analytics, and business intelligence have been headline topics the last couple of years. Similarly, the Internet of Things (IoT) and its many derivatives (Internet of Everything, Internet of People, etc.), have also crashed headlines and are evolving into regular use. One common term, however, has been echoed out of the discussions around all of these terms, and that is machine learning. While you might believe that you are not impacted by many of these technologies, machine learning will affect you.
What Is Machine Learning?
Very simply put, machine learning has to do with data and learning from that data. More specifically, it is a process of making predictions based on patterns, or algorithms, within that data. Unlike data mining that simply looks for information within data, machine learning focuses more on exploring what can be learned from the data. It looks more at the patterns within data rather than simply the information from the data.
Machine learning looks for patterns that can then be applied to do forward predictions. As such, machine learning is equated to predictive analytics. What is important about machine learning is that instead of using static code to make predictions, data is used. This is data that changes, that is dynamic. Because of this dynamic nature, machine learning allows the system to learn and thus evolve with experience and the more data that is analyzed.
Machine Learning in Practice
There are many areas where machine learning will invade your life. Many will be subtle invasions that you might not even notice. Others will begin to be a lot more obvious.
You are reading this article on a website. Data around your usage of websites is being constantly collected. As time goes on, machine learning and predictive analysis will be used to adapt what you are presented with on websites to be more closely suited to what you’ve done in the past. This will help websites customize their content and presentation to what is known about you. More importantly (for them), they will be able to target ads and offers that you are more likely to accept based on what you’ve done in the past.
For this example, it is important to note that the data collected won’t simply be the articles you are reading or the items you’ve clicked. Rather, it will go much deeper and include such tiny details as how much time you spend on a page, how much you scroll, where your mouse pointer resided, what your location was when you were using the site, and much more. It is this massive amount of seemingly trivial data along with the obvious things such as the topics and titles of the content you are reading that will be used to learn about you.
The same type of predictive analysis will be applied across a variety of businesses and industries. Retail stores will get better at not only knowing what to place on shelves, but also at determining what positioning within the store, what colors to use, and what external factors such as lighting also will impact your buying behaviors. Tests in stores such as Target are already being done to learn how to gather the data for systems to learn these behaviors. Those tests include monitoring movements, eye positions, and more, data that will ultimately be used to determine how to create the most effective and engaging displays. When you combine all this data from within the store to the data that can be found in your buying patterns, it will become simple to predicatively sell to you. Remember, the store already knows every item every customer, including you, has ever bought from them with a credit card along with when and where you bought them.
Where is Machine Learning Used Today?
Machine learning can be found in many other areas as well. Within the medical field, machine learning will be used to determine the best diagnosis for a person’s ailments. Machine learning is also what will help natural language processing continue to get better as the amount of voice data continues to increase. Credit card fraud is another area where machine learning is already having an impact. Your buying patterns can be evaluated and predictions made on what you will buy. When something is purchased outside of what you would normally do, credit card companies can flag that. Within robotics, machine learning can help a robot improve its ability to move and interact with what is around it. Robots can adapt to the environment as it encounters more obstacles. Search engines are now using machine learning to help predict what you are going to type into a search box as you are typing. Within gaming, while static rules apply, machine learning can be added to adapt the game to learn how human behaviors occur, thus allowing the machine to predict specifically what you will do in the game versus simply going through a set of static checks that can then be easily countered by a person. The list of examples could go on and on.
Machine Learning Opportunities
Machine Learning generally requires a lot of data combined with serious computing power to do the processing of that data. Detailed data is becoming more and more readily available. With the proliferation of cameras, sensors and IoT, even more data is flooding the market. A lot of this information is already freely available.
For businesses, you should be looking at not only what data is available, but how that data can be used to learn. The examples above simply scratch the surface of what can be done once you start applying algorithms and machine learning.
Recommended Reading: On Developer.com Frank Ohlhorst covers the implications of Machine Learning for Coders.
Machine Language Tools
In the past, the power and tools to grab huge quantities of data and then crunch them into meaningful results was unavailable. Today there are a number of tools already on the market for tapping into machine learning. Many of these tools are provided from within the cloud or from within the major technology companies that you are likely familiar with, such as the following:
- Amazon Machine Learning
- Google TensorFlow
- Microsoft Azure Machine Learning
- IBM SPSS Modeler
These are just a few of the big players. There are a multitude of smaller players too. It is interesting to note, however, that these big players are moving toward giving away the tools they are creating for machine learning. These bigger companies are making their machine learning tools free. This is because machine learning requires data and computational power. Both of these are services that can be bought, especially from these big players!
Machine Learning Still Young
Machine Learning is still young in its lifecycle; however, it is already having an impact. As more data is collected and computing power gets stronger, what can be inferred and learned will grow. This growth is likely to be exponential rather than linear, which means the impact will be felt quicker and quicker. If your company isn’t starting to tap into what can be learned from the relevant data that is available now, then you could quickly be left behind as your competitors get smarter as a result of what they learn from the data. Machine learning is here to stay. It will be interesting to see what it tells us.
Bradley L. Jones is the Director and Editor in Chief of the Developer.com Network of sites, which includes Developer.com, Codeguru, DevX, and HTMLGoodies. He is an internationally bestselling author who has written more than 20 developer-related books across a variety of topics ranging from C++ to Windows and from C# to Web 2.0
This article was originally published on December 02, 2015