What is Machine Learning? A 5 Minute Introduction

Posted by
David Appleyard in Development category

Machine learning still feels like a futuristic concept — a connected set of technologies and algorithms that allow computers to learn new skills, context, or understanding. All without explicit programming.

It explains how Siri learns your voice (and what you’re likely to ask over time), or how Amazon recommends just the right item based on your shopping habits.

Machine learning isn’t just for fun, consumer applications. It can be expanded to the enterprise level for businesses, and have dozens of practical uses such as detecting credit card fraud, helping streamline manufacturing processes, or inventory ordering. Let’s take closer look at what machine learning is, and how it’s likely to present an interesting opportunity for your company.

What Is Machine Learning?

Your first thought might be that machine learning is pretty much the same thing as artificial intelligence. There are lots of parallels, but they’re not quite the same thing. Machine learning is all about big data, modelling, and being able to change and adapt as a program is exposed to new information.

The ideas is that computer programs are able to shift and adapt, using models to “learn” new things over time. Rather than being given a specific set of instructions, in true machine learning, an algorithm has the ability to interpret new information and act upon it based on evolving models. “Learning” could center around audio recognition (e.g. Google Voice or Siri), predicting weather patterns from big data sets, or even tracking typing patterns on a keyboard to determine if the same user is at the controls.

It’s different from data mining in that the computer actually “understands” the information. But like with data mining, a bad or inaccurate set of data can often trip the program up and cause problems.

Exactly how a computer learns has many parallels to how you or I would learn something new. These processes include the use of decision trees, associations, inductive logic, clustering, reinforcement and even forms of “genetic” learning.

Industry Implications

While we’re already seeing machine learning enter various industries in small ways, the technology is still in the very early stages. It has the potential to impact all manner of different industries (and in many ways, already is). Some examples include:

  • Finance: To identify money in/out-flow trends, or patterns to help prevent fraud and other illegal activity.
  • Retail: To suggest products that fit a shopper’s buying habits and are more likely to convert.
  • Transportation: Smart cars are the perfect example of this concept in action, learning routes, hazards and more.
  • Logistics: To better program and direct traffic for deliveries, creating the most efficient route possible.
  • Education: To ensure that students are who they claim to be in online classes and testing environments.

The core of machine learning centres around patterns, and large quantities of data. The exponential increase in computer processing power has lead to opportunities around data analysis and interpretation, making completely new ideas possible.

Whatever industry you work in, there’s a good chance that machine learning is a concept that you’ll want to have a basic understanding of.

Practical Examples

In addition to the broad impact that machine learning will have on entire industries, there are smaller, super-practical applications that you probably already encounter every day. Think about some of the items you might have in your home or be wearing right now:

  • Fitness trackers learn gait and step patterns to help keep track of how much you move (or what type of activity you’re probably doing).
  • Marketing campaigns that adapt based on your needs, preferences, or online activity.
  • Facebook “likes” impact how much content you see from a specific user (or of that specific type of content).
  • Apple/Google Photos learning what the contents of your photos are, to make them searchable and create relevant slideshows/collections.
  • Any type of recommendation system, such as movie suggestions from Netflix.

Where to Learn More

Machine learning is something that can be used or applied to almost any business. Think about your existing data-heavy applications (or the examples above), and how they might relate to opportunities for your company. If you’re interested in learning more, these are all good places to start:

The Evolution of Analytics

This whitepaper takes a look at modern applications for machine learning, with case studies of two companies that have embraced the technology.

A Visual Introduction to Machine Learning

See exactly how a machine learning example comes together, using a real dataset that “knows” the difference between homes in New York and San Francisco.

Coursera’s Machine Learning Class

This courses is taught by a Stanford professor, and designed to provide a broad view of machine learning and the impact it is already having today.

You need to be thinking about machine learning. One way or another, it will impact your business – from the smallest startups to the largest corporations – sooner than you might think. Whether that’s in the technologies you use as part of your business, or in the form of opportunities for new products you could create.

Headquartered in San Francisco, our team of 50+ are fully distributed across 17 countries.