What Is AI, ML, and Deep Learning?
You'll learn to
- -Correctly place AI, machine learning, and deep learning as nested categories, not synonyms
- -Give a precise, working definition of each term
- -Classify a real system by asking two questions: does it learn from data, and does it use a many-layer neural network
- -Say where generative AI, and the large language models in this course, sit in the nesting
Tip: the highlighter is on - just select any text below to mark it. Use the highlighter button up top to change color or turn it off. Highlights are saved on this device only. Log in to save them to your account instead.
These three terms get used interchangeably in casual conversation, which causes real confusion. They are not synonyms. They are nested categories, each one a subset of the one before it.
An Analogy: Boxes Inside Boxes
Think of nested boxes, the kind that fit one inside another. The biggest box is artificial intelligence: any machine doing something we would call intelligent. Inside it is a smaller box, machine learning: the machines that got good at their task by learning from examples. Inside that is a smaller box still, deep learning: machine learning done with many-layered neural networks. Everything in the smallest box is also in the two bigger ones, but not the other way round.
Artificial Intelligence: The Broadest Category
Artificial Intelligence is the broad field of building systems that perform tasks we associate with human intelligence: reasoning, perception, language, decision-making. This includes machine learning, but also things that involve no learning at all. A chess engine built entirely from hand-coded search and evaluation rules is AI, even though it never learns from data.
Machine Learning: Learning From Data
Machine Learning is the subset of AI where the system improves its behavior by learning patterns from data, rather than following only hand-written rules. A spam filter trained on labeled emails is machine learning. A chess engine using only hard-coded heuristics is not, even though both are AI.
Deep Learning: Learning With Neural Networks
Deep Learning is the subset of machine learning that uses neural networks with many layers, "deep" refers to layer count, to learn patterns directly from raw or lightly processed data such as pixels, audio waveforms, or raw text, instead of requiring a human to hand-engineer features first. Large language models are deep learning systems.
One Task, Three Approaches
The clearest way to see the difference is to solve the same problem three ways. Take spam filtering:
- -Hand-written rules: if the email contains free money, mark it as spam. A programmer wrote every rule and nothing is learned. This is AI, but not machine learning.
- -A model trained on labeled emails: a program counts which words appear in spam and in normal mail, and learns how much each word should count. A person chose the features (word counts), and the machine learned the weights. This is machine learning, but not deep learning.
- -A deep neural network that reads the raw text of millions of emails and works out for itself which patterns matter, with no one choosing the features. This is deep learning.
Notice the direction of progress: each step takes more of the work out of the programmer's hands and gives it to the data. Deep learning's real novelty is that the machine learns the features too, which is why it works on messy raw inputs such as pixels, sound and text.
Where This Course Is Heading
The systems this course is about, large language models and the image generators you may have seen, are built with deep learning. They belong to a family called generative AI: systems that produce new content, such as text or pictures, instead of only labeling or predicting. So generative AI, as it is practiced today, sits inside deep learning, which sits inside machine learning, which sits inside AI.
A common mistake is calling every AI system "deep learning." A simple linear regression model predicting house prices is machine learning, but it is not deep learning. There is no neural network involved.
A job posting asks for "deep learning experience" for a role that turns out to be building a simple linear regression model for pricing. Is that a red flag?
"No, deep learning and ML are basically the same thing, so it does not matter."
"It is a signal worth asking about, since deep learning specifically means neural networks with many layers, and a linear regression model has no neural network in it at all, it is classical ML. Either the posting is using the terms loosely, which is common and not necessarily disqualifying, or there is a real mismatch between what the role advertises and what it actually needs. I would ask directly what modeling techniques the team uses day to day, since the answer tells you a lot about whether the posting is accurate."
What is the correct relationship between AI, machine learning, and deep learning?
Discussion0
Join the Discussion
Sign in to leave comments, reply to others, or like insights.
No comments yet. Be the first to start the thread!