AI & ML Orientation

What AI and ML are, how learning differs from rules, and how to choose a first path.

For a small example, sort three support messages with a keyword rule, then with labeled examples. The rule is transparent but brittle; the learned boundary handles variation only when the examples represent it.

In practice

A clinic wants to prioritize appointment reminders for likely no-shows. Frame a prediction, define the intervention, compare with a simple baseline, and keep staff responsible for the final action.

Initializing Interactive Playground...

Experiment notes

Record your experiment

Before opening the experiment, predict whether changing the observed hours or the threshold changes the rule, the evidence, or both, and identify the first unfair comparison to avoid.

PredictionWhat changed, and why?

Implementation

Build the transparent baseline first

A team wants a neural network for a tiny pass-risk example but has not implemented the threshold rule or calculated its accuracy.

Run the starter and state the prediction unit, target, and available input.
Implement the threshold rule and calculate correct predictions divided by total labels.
Explain what protected evidence a learned candidate must beat and which error cost accuracy hides.

Check: The checks pass, the baseline is reproducible, and your explanation separates a working rule from evidence that machine learning is warranted.

Failing starter lab

Compare a rule with observed labels

Complete a tiny threshold classifier and compute its accuracy. The point is to establish a transparent baseline before considering a learned model.

Run the incomplete starter before opening hints. Use the first wrong output or failed check to form a repair hypothesis.

Console

Run the code to inspect its output. Use Check solution only when you want to run the exercise checks.

Code runs only in a time-limited browser worker. It is not sent to the IntuitiveAI server. Do not paste secrets.

Expected result

predictions: [0, 0, 0, 1, 1, 1]
accuracy: 1.0
ALL TESTS PASSED

Chapter 1 of 4 · Start with the map

A useful map of AI and machine learning

Artificial intelligence is the broad goal of building systems that perform tasks associated with reasoning, perception, language, or decision-making. Machine learning is one way to build such systems: instead of specifying every rule, we fit behavior from examples or feedback. Deep learning is a family of machine-learning methods built from layered neural networks. Generative AI produces new content by modeling patterns in training data.

These categories overlap, but they are not synonyms. A tax calculator can be useful automation without ML. A spam filter can be ML without deep learning. A language model is deep learning and generative AI, but the surrounding product also contains deterministic software, retrieval, safety controls, and monitoring.

Ideas to understand

AI is the broad system goal; ML is one implementation strategy.
Training estimates model parameters from data; inference applies the fitted model.
A model is only one component of an AI product.
The simplest adequate non-ML baseline is valuable evidence.

NextTurn a product question into a learning problem

Video explanation

AI, Machine Learning, Deep Learning and Generative AI Explained

IBM Technology

Separate the major AI terms before deciding which kind of system or learning problem you are working on.

Watch on YouTube

References

Primary sources and free technical reading for AI & ML Orientation.

The Python Tutorial

The official, free introduction to Python syntax and core programming ideas.

DocumentationFree

Google Machine Learning Glossary

A free glossary for core ML terminology and common model concepts.

DocumentationFree

Back to Topic Library

Feedback & sharing

AI & ML Orientation