If you’re new to the field of data science, it may seem like there’s a lot of jargon to keep track of. Two of the loudest buzz words, which are often mistakenly used interchangeably, are machine learning and deep learning. What’s the difference? The short answer is that deep learning is a technique for implementing machine learning. But let’s zoom out for a minute and discuss what both of these terms mean on a larger scale, and how they fit into the larger scope of artificial intelligence.
Artificial Intelligence: The Outer Circle
Draw three concentric circles, and you will have a helpful visual aid for thinking about how artificial intelligence, machine learning, and deep learning relate to each other.
Artificial intelligence, or the outermost circle, can be defined as the capability of a machine to imitate intelligent human behavior. This includes the capacity to handle complicated tasks such as decision making, understanding human speech, detecting fraud, and more. As Netflix’s Senior Data Scientist F. William High said during the opening panel at DataScience: Elevate, “AI implies a level of system control and orchestration of multiple models and rules."
We can think of machine learning as an important subset of AI, encompassing the techniques and strategies that work to answer the questions that AI is trying to answer.
Machine Learning: The Middle Circle
Machine learning is defined by Stanford University as the science of getting computers to act in specific ways without explicitly programming them to do so. Why is this important? Data is rapidly accumulating by the second. In order to keep up, it’s essential to implement algorithms that can quickly and efficiently recognize patterns and make predictions.
There are a number of machine learning methods or algorithms that can be applied to almost any data problem. These techniques include linear regression, logistic regression, k-means clustering, decision trees, random forests, and more, and they can all be applied to a variety of real-life use cases. Some of these use cases were discussed during DataScience: Elevate's opening panel, and they range from building recommendation engines for curated email content at Quora, to employing natural language processing in chat logs at Riot Games, to building predictive models focused on customer churn at Verizon Wireless.
Deep Learning: The Inner Circle
Deep learning is a form of machine learning that is inspired by the structure of the human brain and is particularly effective in feature detection.
This technique involves feeding your model large volumes of data, but it requires less feature engineering than a linear regression model would. How does this translate to real life? "If you’re looking at classifying images for a cat, you’ll have to feed your data set with a bunch of images of a cat. But you don’t necessarily have to say that a cat is something with cute ears or whiskers,” explained Verizon Wireless Data Scientist Aurora LePort during the opening panel of DataScience: Elevate.
Before you start thinking about deep learning, however, it’s important to first fully understand the concept of a neural network. A neural network passes data through interconnected layers of nodes, classifying information and characteristics of a layer before passing the results on to other nodes in subsequent layers. The difference between a neural network and a deep learning network is contingent on the number of layers: A basic neural network may have two to three layers, while a deep learning network may have dozens or hundreds.
The technique of deep learning has garnered a tremendous amount of buzz, particularly because of how it uses the human brain and neural coding as the basis for how a machine can recognize and classify stimuli.
Hopefully this post gives you a helpful frame of reference for thinking about the differences between artificial intelligence, machine learning, and deep learning, as well as the ways that they are all connected.