Machine Learning: The Basics

Stephen DeAngelis

June 25, 2020

Machine learning (ML) is one of the basic technologies found in the field of artificial intelligence (AI). The two terms are often used interchangeably, but they are not the same thing — machine learning is a subset of artificial intelligence. Some pundits insist AI remains an aspirational concept, while machine learning is being used widely by the business community. The staff at the Washington University Olin Business School asserts, “The words ‘machine learning’ are more than the latest business buzz words. If you’re not tapping into the power of big data analytics, you’re already behind the curve.”[1] Most businesses strive to get ahead of the game so being behind the curve can be unnerving. The Olin Business School staff adds, “Businesses in every industry are unearthing profit-generating insights unknowable just a decade ago, with customer data collected from scanners, cash registers, online product reviews, wearable devices and other sources.”

The basics

Noah Heinrich writes, “Machine learning does exactly what it says on the tin. It is a method by which a computer program can ‘automatically learn and improve from experience without being explicitly programmed.’ … In broad strokes, a computer program using machine learning … analyzes data and searches for an underlying pattern or trend to develop a predictive model that learns from the data it’s fed.”[2] Heinrich’s last point is important to remember — there is no machine learning without data. The fact that machines learn from data also highlights why machine learning can’t be equated to artificial intelligence. Alan J. Porter (@alanjporter), Head of Strategic Services for [A], explains, “ML is great at recognizing patterns but not much else” and “ML assumes tomorrow is going to be the same as today.”[3] For many business applications, that’s good enough. For other business challenges, it’s not nearly enough.

Computers (aka machines) need programs to direct their learning efforts. Those programs are called algorithms. Heinrich explains, “Machine learning algorithms can process massive amounts of data and predict outcomes and patterns based on that information. Over time, the predictive model becomes more accurate as the program improves itself, no outside tampering required. There are three broad categories of algorithms, which are defined by what kind of training datasets they are given: supervised, unsupervised, and semi-supervised. Each of these approaches has advantages and disadvantages, depending on what the program is intended to accomplish.” Cynthia Harvey explains there are actually four, not three, types of machine learning.[4] They are:

  • Supervised Learning: “Supervised learning requires a programmer or teacher who offers examples of which inputs line up with which outputs.”
  • Unsupervised Learning: “Unsupervised learning requires the system to develop its own conclusions from a given data set.”
  • Semi-supervised Learning: “Semi-supervised learning, as you probably guessed, is a combination of supervised and unsupervised learning.”
  • Reinforcement Learning: “Reinforcement learning involves a system receiving feedback analogous to punishments and rewards.”

Dr. Marc Deisenroth (@mpd37), a lecturer at Imperial College London, adds, “Machine learning can be considered the engine of modern AI. It provides the underlying technology that drives AI. AI is about complex systems that behave intelligently. In order to reach this goal, AI poses many questions, and machine learning provides the technologies toward answering these questions. In other words, AI is about systems and questions whereas machine learning is about practical solutions to these challenges. Another difference is that AI strives for intelligence, whereas machine learning does not necessarily do this.”[4]

Choosing the right machine learning model

Machine learning’s goal is to perform either a classification or a regression. Classification involves determining groupings, such as determining the breed of a dog within a picture. Regression involves estimating numeric values, such as projecting sales. Both cases involve independent variables (inputs) and dependent variables (outputs being predicted). There is no single “business algorithm” that addresses the numerous challenges organizations confront. To extract value using analytics you need the right data and the right model. Daniel Tunkelang (@dtunkelang), who led machine learning projects at Endeca, Google, and LinkedIn, explains, “There are an incredible number of problems that you can solve by providing the right training data to the right learning algorithms. … [However,] machine learning isn’t as simple as ‘if you build it, it will learn.’ You need to make sure you’re giving it the right input and that you’re applying the right model. Your data is your cake mix, and your model is your method. Get it right, and you can have your cake and eat it, too. Get it wrong, and you’ll have to start over.”[6] To ensure the right model is used for our clients, my company, Enterra Solutions®, leverages the Representational Learning Machine™ (RLM) created by Massive Dynamics™. The RLM can help determine what type of analysis is best-suited for the data involved in a high-dimensional environment. Obtain the right data, identify a specific problem, and apply the right analytics model and you get the results for which you are looking. It’s not hype, but it’s not as easy as it sounds. Alan Descoins (@dekked_), CTO & partner at Tryolabs, suggests executives ask eleven questions before starting any machine learning project.[7] Those questions are:

1. What are your organization’s business goals?
2. Should machine learning reduce costs or increase revenue?
3. Which is your clear and realistic way of measuring the success of your ML initiative?
4. How does your organization handle the risk?
5. How do you acquire the right talent?
6. Do you have a clear high-level understanding of what machine learning is?
7. Is access to information guaranteed?
8. Have you planned the initiative as a mid-term project?
9. Is your organization collecting the right data?
10. Is your organization collecting the data in the right format?
11. Have you taken human labeling of data into consideration?

For a fuller discussion of each question, read Descoins’ article. He concludes, “Machine learning algorithms are changing nearly every industry. They’re increasing productivity, boosting sales and helping us make more informed decisions.”

Concluding thoughts

The question is not “if’ companies will leverage machine learning in the future but “how” they will leverage it. Dr. Nipa Basu (@nipabasu), Chief Analytics Officer at Dun & Bradstreet, concludes, “Machine learning is changing the way businesses look at data and presenting new analytics opportunities for companies of all sizes. Increasingly, how organizations leverage new technology for machine learning in business will be a key deciding factor in whether they can ride the waves of change or find themselves washed up on the data analytics shore.”[8]

Footnotes
[1] Washington University Olin Business School, “Got algorithm? Machine learning quickly evolving to solve business problems,” Minneapolis/St. Paul Business Journal, 27 February 2020.
[2] Noah Heinrich, “What is Machine Learning? A Quick Guide to Basic Concepts,” Built In, 8 January 2020.
[3] Alan J. Porter, “Machine Learning Isn’t Rocket Science,” CMS Wire, 9 September 2019.
[4] Cynthia Harvey, “What is Machine Learning?” Datamation, 3 January 2018.
[5] Colin Smith, “Machine learning research: the driving force of Artificial Intelligence,” Imperial College London, 4 October 2017.
[6] Daniel Tunkelang, “Ten Things Everyone Should Know About Machine Learning,” Forbes, 6 September 2017.
[7] Alan Descoins, “11 questions to ask before starting a successful machine learning project,” Customer Think, 25 February 2019.
[8] Nipa Basu, “3 ways to make machine learning in business more effective,” TechTarget, October 2018.