YOUR FOUNDATION-FIRST SYLLABUS
Prediction & anomaly detection
Use past examples to predict an amount or a category and notice unusual cases. Compare simple methods, test on examples kept out of training, and investigate different kinds of mistakes.
0 of 45 core ideas demonstrated45 lessons595 estimated minutes remaining
Already know some of this?
Go straight to an idea’s knowledge check. Passing its different checks without hints carries that evidence into this path. Checking an advanced idea does not award its prerequisites; you can still explore them.
Check what I already knowTHE IDEAS, IN ORDER
Shared credit is automaticBuild your understanding.
Chapter counts use compatible knowledge-check evidence. Reading a lesson does not mark its ideas as demonstrated.
Chapter 1 · Lessons 1–7 Meet an AI system Start with an everyday task, then follow its inputs, model, and output. 7 lessons · 0 of 7 core ideas demonstrated Up next
01
What is AI? Start with an everyday task Common Ground · 11 min
To explore02 Inputs and outputs: what goes in and what comes out Common Ground · 13 min
To explore03 Data and examples: what a computer can learn from Common Ground · 10 min
To explore04 What is a model? A small rule inside a bigger app Common Ground · 10 min
To explore05 Weights: the adjustable numbers in a model Common Ground · 12 min
To explore06 Training and inference: changing a rule or using it Common Ground · 13 min
To explore07 Follow one small AI system from start to finish Common Ground · 9 min
To exploreChapter 2 · Lessons 8–14 How computers represent a problem Explore everyday AI ideas; how AI got here; how computers represent a problem; and supporting ideas. 7 lessons · 0 of 7 core ideas demonstrated
08
Different ways to solve the same problem Common Ground · 11 min
To explore09 How to tell whether an AI answer is useful Common Ground · 10 min
To explore10 The questions that shaped AI Common Ground · 15 min
To explore11 How a computer stores a picture or a sentence Learning mechanisms · 16 min
To explore12 Describe a computation so someone else can follow it Learning mechanisms · 11 min
To explore13 What a representation makes easy to learn Learning mechanisms · 12 min
To explore14 From the world to a dataset Learning mechanisms · 11 min
To exploreChapter 3 · Lessons 15–21 Learning to predict: Build a tiny prediction model Explore learning to predict; math preparation when it is needed. 7 lessons · 0 of 7 core ideas demonstrated
15
The numbers you actually need Math preparation · 10 min
To explore16 A function is a rule Math preparation · 9 min
To explore17 Build a tiny prediction model Learning mechanisms · 11 min
To explore18 Vectors: lists that work together Math preparation · 10 min
To explore19 Distances, neighborhoods, and transformations Math preparation · 12 min
To explore20 Predict from similar stored examples Learning mechanisms · 11 min
To explore21 Probability without the mystery Math preparation · 10 min
To exploreChapter 4 · Lessons 22–29 Learning to predict: Split a dataset into simpler regions Explore learning to predict; learning from useful data; finding patterns and useful representations; and supporting ideas. 8 lessons · 0 of 8 core ideas demonstrated
22
Split a dataset into simpler regions Learning mechanisms · 12 min
To explore23 Powers and logarithms, one step at a time Math preparation · 11 min
To explore24 Turn a linear score into a class probability Learning mechanisms · 11 min
To explore25 Averages over uncertain outcomes Math preparation · 9 min
To explore26 Who is represented by the data? Learning mechanisms · 14 min
To explore27 Decide what success means before training Learning mechanisms · 12 min
To explore28 Look inside the average score Learning mechanisms · 14 min
To explore29 Find groups that help with your task Learning mechanisms · 17 min
To exploreChapter 5 · Lessons 30–37 Working with uncertainty Explore working with uncertainty; finding patterns and useful representations; learning from useful data; and supporting ideas. 8 lessons · 0 of 8 core ideas demonstrated
30
Probability, likelihood, and what is unknown Learning mechanisms · 11 min
To explore31 Find unusual cases worth investigating Learning mechanisms · 13 min
To explore32 What can a small experiment tell us? Math preparation · 10 min
To explore33 Clean data without erasing the problem Learning mechanisms · 12 min
To explore34 Protect the examples used to judge a model Learning mechanisms · 15 min
To explore35 Control complexity without peeking at the answer Learning mechanisms · 12 min
To explore36 Evaluate the procedure that chose the model Learning mechanisms · 14 min
To explore37 Update a belief using new evidence Learning mechanisms · 13 min
To exploreChapter 6 · Lessons 38–45 Testing what a model has learned Explore working with uncertainty; testing what a model has learned; how computers represent a problem; and supporting ideas. 8 lessons · 0 of 8 core ideas demonstrated
38
Value a choice and its later consequences Math preparation · 12 min
To explore39 Turn a probability into a justified action Learning mechanisms · 11 min
To explore40 Change one thing and measure what follows Learning mechanisms · 13 min
To explore41 Repeat an experiment without confusing luck with truth Learning mechanisms · 11 min
To explore42 Make the result possible to inspect Learning mechanisms · 12 min
To explore43 Measure the behavior that the task actually needs Learning mechanisms · 14 min
To explore44 Decide whose problem the system solves and who bears its errors Learning mechanisms · 12 min
To explore45 Project: compare a baseline with a learned predictor Learning mechanisms · 75 min
To exploreOptional extensions and comparisons (19)
These lessons deepen or compare the reference design. They do not add requirements to this path’s completion.
How to use this school without getting lostOptional orientation: how paths, side lessons, and practice fit together, plus study habits that make learning stick.Practice, progress, and knowing what you understandOptional orientation: use practice, mistakes, and spaced review to find out what you really understand.Choose an answer from a listFast typed decisions for routing and triage: when a small decision model beats a chat model.Combine many imperfect predictorsRandom forests and gradient boosting, still the champions on spreadsheet-style data.Choose a boundary with a marginThe maximum-margin classifier that dominated machine learning before deep learning.Ask what would change if an action changed the worldPrediction isn't intervention. Learn when a model can tell you what would happen if you acted.Notice when the production task changesYour model worked at launch. Learn to notice when the world changes under it.Create more examples without inventing evidenceStretch a small dataset with transformed or generated examples without fooling yourself.A useful classifier with a strong independence assumptionThe classic spam filter: a fast baseline built on one bold simplifying assumption.Let several hidden sources explain the dataExplain data as a blend of hidden groups, and meet the EM algorithm.Read a data map with careRead t-SNE and UMAP plots without being misled by the pretty clusters.What can a learning guarantee actually guarantee?What can math actually promise about how a model will do on new data?State the assumptions behind effects that cannot all be observedWhat would have happened otherwise? The assumptions that let data answer 'what if' questions.Distinguish examples in a prompt from learning to adapt weightsLearn from a handful of examples, including how language models do it inside a prompt.Choose which examples to learn from nextLabels are expensive. Learn to pick the examples that teach a model the most per label.Use unlabeled data and imperfect label sources carefullyLearn from mostly unlabeled data and noisy labels without fooling yourself.Coordinate learning while data remain distributedTrain on data that never leaves people's phones, and see what privacy that does and doesn't buy.Give people useful control over model-assisted workDesign the review step so people catch model mistakes instead of rubber-stamping them.Apply the same lifecycle questions to different productsSee how the same lifecycle plays out in recommendation, forecasting, search, and accessibility products.