Even if you’re not involved in data science, you’ve probably heard the terms artificial intelligence (AI), machine learning (ML), and deep learning in recent years. In some cases, they may be used interchangeably. Although these terms are related, each has distinct meanings and is more than just a buzzword used to describe self-driving cars.
In a broader sense, deep learning is a subset of machine learning, and machine learning is a subset of artificial intelligence. You can think of these as a series of overlapping concentric circles, with AI leading the way, followed by machine learning and deep learning. In other words, deep learning is AI, but AI is not deep learning.
Learn more about AI, machine learning, and deep learning, including how they are related and different.
Deep learning and machine learning
The Oxford English Dictionary defines AI as “the ability of a computer or other machine to exhibit or simulate intelligent behavior.” [1]. Britannica offers a similar definition. “The ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.” [2].
Machine learning and deep learning are both types of AI. In other words, machine learning is AI that can automatically adapt with minimal human intervention. Deep learning is a subset of machine learning that uses artificial neural networks to mimic the learning process of the human brain.
Look at these key differences.
![[Diagram] A Venn diagram on a blue background showing how deep learning, machine learning, and AI are nested.](https://images.ctfassets.net/wp1lcwdav1p1/7aSfuCL24ZdzooNkTcZd90/4712a602f1e248b794c1d3631301bef0/image1.png?w=1500&q=60)
| machine learning | deep learning |
|---|---|
| A subset of AI | Machine learning subset |
| Can be trained on smaller datasets | Requires large amounts of data |
| Requires more human intervention to collect and learn | Able to learn from the environment and past mistakes |
| Training is shorter and accuracy is lower | Longer training and higher accuracy |
| It usually creates a simpler, more linear correlation. | Creates non-linear and complex correlations |
| Can be trained on a central processing unit (CPU) | Training requires a special graphics processing unit (GPU) |
What is artificial intelligence (AI)?
At the most basic level, the field of artificial intelligence uses computer science and data to enable machines to solve problems.
Although we don’t yet have human-like robots trying to conquer the world, there are examples of AI all around us. These can be as simple as a computer program that can play chess, or as complex as an algorithm that can predict the ribonucleic acid (RNA) structure of a virus to help develop a vaccine.
Machine learning is needed to independently improve machines and programs without it Further input from human programmers.
Deep Blue, a computer playing chess
Before the development of machine learning, artificial intelligence machines or programs had to be programmed to respond to a limited set of inputs. Deep Blue, the chess-playing computer that defeated the world chess champion in 1997, was able to “determine” the next move based on an extensive library of possible moves and outcomes. [3].
What is machine learning?
Machine learning refers to the study of computer systems that automatically learn and adapt from experience without being explicitly programmed by humans.
With simple AI, programmers can tell machines how to respond to different commands by hand-coding each “decision.” Machine learning models allow computer scientists to “train” machines by feeding them large amounts of data. Machines follow a set of rules called algorithms to analyze data and draw inferences from it. The more data a machine analyzes, the better it can perform tasks and make decisions.
For example, the music streaming service Spotify learns about your music preferences and provides new suggestions. Each time you indicate that you like a song by listening to it all the way through or adding it to your library, the service updates its algorithm to provide more accurate recommendations. Netflix and Amazon use similar machine learning algorithms to provide personalized recommendations.
What is deep learning?
While machine learning algorithms typically require human correction if something goes wrong, deep learning algorithms can iterate to improve results without human intervention. While machine learning algorithms can learn from relatively small data sets, deep learning algorithms require large data sets that can include diverse unstructured data.
Think of deep learning as an evolution of machine learning. Deep learning is a machine learning technique that layers algorithms and computing units, or neurons, into artificial neural networks. These deep neural networks are inspired by the structure of the human brain. Data passes through this web of interconnected algorithms non-linearly, similar to how our brains process information.
What’s wrong with big data?
“Big data” refers to data sets that are too large to be managed by traditional relational databases and data processing software. Businesses are generating unprecedented amounts of data every day, and deep learning is one way to extract value from that data.
Learn more about deep learning, machine learning, and AI on Coursera
AI, machine learning, and deep learning are all connected. Deep learning is the most advanced and requires large amounts of data to learn and create complex nonlinear correlations. Machine learning is simpler, requires less data, and yields linear correlations.
Continue exploring technology with programs available on Coursera. For example, if this introduction to AI, deep learning, and machine learning piques your interest, consider taking AI forEveryone, a course designed to teach the basics of AI to students from non-technical backgrounds.
If you need more advanced knowledge, start with Andrew Ng’s Machine Learning Specialization, which provides a comprehensive introduction to machine learning concepts. Next, build and train an artificial neural network with the deep learning specialization.
When you’re ready, earn the IBM Data Science Professional Certificate and start building the skills you need for an entry-level role as a data scientist.
