Page 01 of 14 · Foundations
What is AI, and where does machine learning fit?
For most of computing history, a person wrote every rule a program followed. Machine learning flips that around: you give the computer examples, and it works out the rules for itself.
You write if subject contains "FREE" then spam. Spammers write "FR3E" and your rule breaks.
You give it 50,000 emails already marked spam or not spam. It finds the patterns, including ones you never thought to look for.
The learned "rules" are called a model. Once trained, the model answers new questions it has never seen.
AI, ML, and deep learning are nested
Select a ring to see what belongs inside it.
People use these three terms as if they mean the same thing. They don't. Each one is a subset of the one around it, so all deep learning is machine learning, and all machine learning is AI, but not the other way around.
Words you will see on every page
Select a term to see its definition.
Most of this course fits in one line of notation. A model f with parameters θ takes features x and predicts a label:
Read it as "y-hat equals f of x." The hat means "predicted," so ŷ is the model's guess and y is the true answer. Training adjusts θ to shrink the gap between them.
Three ways a machine can learn
Machine learning splits by what kind of feedback the model gets during training.
Every example comes with the right answer (a label). The model learns to predict the label. Predicting a number is regression; predicting a category is classification.
No labels at all. The model looks for structure on its own, such as groups of similar items or points that don't fit anywhere.
An agent takes actions and gets a reward or penalty. Over many tries it learns which actions pay off.
Which kind of learning?
Read each scenario and choose the type of learning it describes.
How we got here
AI is older than most people think. The ideas behind today's neural networks date to the 1950s; what changed recently is the amount of data and computing power available to train them.
Why this matters in cybersecurity
Machine learning already sits inside many of the tools you will use: spam and phishing filters, intrusion detection systems that flag unusual traffic, malware classifiers, and fraud detection on payment networks. Attackers use it too, to write convincing phishing messages and to find weaknesses faster.