AI-02 · Undergraduate · ENGLISH · PLANNED
Begin with a baseline, trace a model, measure errors and improve it.
Study time is an initial activity estimate, excluding extra reading, revision and extended laboratory or project work. Videos and question bodies have not been produced by adding this outline.
Complete these courses or demonstrate equivalent knowledge. These are recommended prerequisites, not enforced enrolment locks.
YouTube playlist: ARK Artificial Intelligence | Mathematics and Data for AI
Playlist title reserved in the plan; no YouTube playlist or upload has been created yet.
Sign in to save this learning pathPractical outcome: Trace the shape and gradient of a tiny prediction model.
Explain vectors as features using a visual and one worked real-world example. Context: Trace the shape and gradient of a tiny prediction model. End with a prediction or check-for-understanding.
Explain matrix batches using a visual and one worked real-world example. Context: Trace the shape and gradient of a tiny prediction model. End with a prediction or check-for-understanding.
Explain dot-product scores using a visual and one worked real-world example. Context: Trace the shape and gradient of a tiny prediction model. End with a prediction or check-for-understanding.
Explain matrix shape checking using a visual and one worked real-world example. Context: Trace the shape and gradient of a tiny prediction model. End with a prediction or check-for-understanding.
Use the first concepts from this section. Trace the shape and gradient of a tiny prediction model. Produce a labelled diagram, trace or calculation; explain each step and check one changed input.
3–6 questions on the preceding first-half concepts only: recall, visual interpretation, application and one plausible misconception. Explain each answer after submission.
Explain derivatives of losses using a visual and one worked real-world example. Context: Trace the shape and gradient of a tiny prediction model. End with a prediction or check-for-understanding.
Explain partial gradients using a visual and one worked real-world example. Context: Trace the shape and gradient of a tiny prediction model. End with a prediction or check-for-understanding.
Explain chain rule in networks using a visual and one worked real-world example. Context: Trace the shape and gradient of a tiny prediction model. End with a prediction or check-for-understanding.
Explain gradient checking using a visual and one worked real-world example. Context: Trace the shape and gradient of a tiny prediction model. End with a prediction or check-for-understanding.
Trace the shape and gradient of a tiny prediction model. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Vectors as features, Matrix batches, Dot-product scores and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Practical outcome: Compare two classifiers using conditional probabilities.
Explain conditional models using a visual and one worked real-world example. Context: Compare two classifiers using conditional probabilities. End with a prediction or check-for-understanding.
Explain likelihood using a visual and one worked real-world example. Context: Compare two classifiers using conditional probabilities. End with a prediction or check-for-understanding.
Explain log likelihood using a visual and one worked real-world example. Context: Compare two classifiers using conditional probabilities. End with a prediction or check-for-understanding.
Explain bayes' rule for prediction using a visual and one worked real-world example. Context: Compare two classifiers using conditional probabilities. End with a prediction or check-for-understanding.
Use the first concepts from this section. Compare two classifiers using conditional probabilities. Produce a labelled diagram, trace or calculation; explain each step and check one changed input.
3–6 questions on the preceding first-half concepts only: recall, visual interpretation, application and one plausible misconception. Explain each answer after submission.
Explain entropy using a visual and one worked real-world example. Context: Compare two classifiers using conditional probabilities. End with a prediction or check-for-understanding.
Explain cross entropy using a visual and one worked real-world example. Context: Compare two classifiers using conditional probabilities. End with a prediction or check-for-understanding.
Explain kl divergence using a visual and one worked real-world example. Context: Compare two classifiers using conditional probabilities. End with a prediction or check-for-understanding.
Explain sampling distributions using a visual and one worked real-world example. Context: Compare two classifiers using conditional probabilities. End with a prediction or check-for-understanding.
Compare two classifiers using conditional probabilities. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Conditional models, Likelihood, Log likelihood and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Practical outcome: Build a reproducible synthetic-data split without leakage.
Explain target definition using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Explain data provenance using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Explain train-validation-test splits using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Explain missing data using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Use the first concepts from this section. Build a reproducible synthetic-data split without leakage. Produce a labelled diagram, trace or calculation; explain each step and check one changed input.
3–6 questions on the preceding first-half concepts only: recall, visual interpretation, application and one plausible misconception. Explain each answer after submission.
Explain scaling using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Explain categorical encoding using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Explain leakage using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Explain class imbalance using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Explain baseline models using a visual and one worked real-world example. Context: Build a reproducible synthetic-data split without leakage. End with a prediction or check-for-understanding.
Build a reproducible synthetic-data split without leakage. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Target definition, Data provenance, Train-validation-test splits and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Combine all three sections: Build a reproducible synthetic-data split without leakage. Submit an original worked artifact, evidence or tests, assumptions and a reflection. Rubric: correctness, method, practical usefulness, reproducibility and communication. Self-reviewed, not certified.
6–12 balanced questions sampling every section of Mathematics and Data for AI, including interpretation, calculation or reasoning, a misconception and a new application. Passing is 70%; review errors before retrying.
Data experiments, evaluations, model tracing and reproducible projects. Guided work leads to an independent attempt, with a diagram, explanation, trace, calculation or tested artifact. Question bodies are generated when opened in your private learning path. Programming submissions are stored as text; the platform does not execute submitted code.
Use original or appropriately licensed datasets. Practical science activities use models or supplied data unless a qualified supervisor has reviewed a physical experiment.
Open the interactive practice sampler →