AI-13 · Advanced · 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 | Responsible AI, Causality and Scientific Evaluation
Playlist title reserved in the plan; no YouTube playlist or upload has been created yet.
Sign in to save this learning pathPractical outcome: Distinguish a predictive association from an intervention claim.
Explain confounding using a visual and one worked real-world example. Context: Distinguish a predictive association from an intervention claim. End with a prediction or check-for-understanding.
Explain causal graphs using a visual and one worked real-world example. Context: Distinguish a predictive association from an intervention claim. End with a prediction or check-for-understanding.
Explain interventions using a visual and one worked real-world example. Context: Distinguish a predictive association from an intervention claim. End with a prediction or check-for-understanding.
Explain counterfactuals using a visual and one worked real-world example. Context: Distinguish a predictive association from an intervention claim. End with a prediction or check-for-understanding.
Use the first concepts from this section. Distinguish a predictive association from an intervention claim. 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 identification assumptions using a visual and one worked real-world example. Context: Distinguish a predictive association from an intervention claim. End with a prediction or check-for-understanding.
Explain randomised experiments using a visual and one worked real-world example. Context: Distinguish a predictive association from an intervention claim. End with a prediction or check-for-understanding.
Explain matching concepts using a visual and one worked real-world example. Context: Distinguish a predictive association from an intervention claim. End with a prediction or check-for-understanding.
Explain sensitivity analysis using a visual and one worked real-world example. Context: Distinguish a predictive association from an intervention claim. End with a prediction or check-for-understanding.
Distinguish a predictive association from an intervention claim. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Confounding, Causal graphs, Interventions and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Practical outcome: Evaluate a fictional selection model using multiple affected-user perspectives.
Explain fairness definitions using a visual and one worked real-world example. Context: Evaluate a fictional selection model using multiple affected-user perspectives. End with a prediction or check-for-understanding.
Explain metric tradeoffs using a visual and one worked real-world example. Context: Evaluate a fictional selection model using multiple affected-user perspectives. End with a prediction or check-for-understanding.
Explain representational harms using a visual and one worked real-world example. Context: Evaluate a fictional selection model using multiple affected-user perspectives. End with a prediction or check-for-understanding.
Explain privacy leakage using a visual and one worked real-world example. Context: Evaluate a fictional selection model using multiple affected-user perspectives. End with a prediction or check-for-understanding.
Use the first concepts from this section. Evaluate a fictional selection model using multiple affected-user perspectives. 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 differential privacy concepts using a visual and one worked real-world example. Context: Evaluate a fictional selection model using multiple affected-user perspectives. End with a prediction or check-for-understanding.
Explain explainability limits using a visual and one worked real-world example. Context: Evaluate a fictional selection model using multiple affected-user perspectives. End with a prediction or check-for-understanding.
Explain model documentation using a visual and one worked real-world example. Context: Evaluate a fictional selection model using multiple affected-user perspectives. End with a prediction or check-for-understanding.
Explain human oversight using a visual and one worked real-world example. Context: Evaluate a fictional selection model using multiple affected-user perspectives. End with a prediction or check-for-understanding.
Evaluate a fictional selection model using multiple affected-user perspectives. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Fairness definitions, Metric tradeoffs, Representational harms and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Practical outcome: Audit a model claim with baselines uncertainty and failure analysis.
Explain benchmark design using a visual and one worked real-world example. Context: Audit a model claim with baselines uncertainty and failure analysis. End with a prediction or check-for-understanding.
Explain distribution shift using a visual and one worked real-world example. Context: Audit a model claim with baselines uncertainty and failure analysis. End with a prediction or check-for-understanding.
Explain robustness using a visual and one worked real-world example. Context: Audit a model claim with baselines uncertainty and failure analysis. End with a prediction or check-for-understanding.
Explain calibration assessment using a visual and one worked real-world example. Context: Audit a model claim with baselines uncertainty and failure analysis. End with a prediction or check-for-understanding.
Use the first concepts from this section. Audit a model claim with baselines uncertainty and failure analysis. 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 uncertainty estimation using a visual and one worked real-world example. Context: Audit a model claim with baselines uncertainty and failure analysis. End with a prediction or check-for-understanding.
Explain reproducibility using a visual and one worked real-world example. Context: Audit a model claim with baselines uncertainty and failure analysis. End with a prediction or check-for-understanding.
Explain statistical comparisons using a visual and one worked real-world example. Context: Audit a model claim with baselines uncertainty and failure analysis. End with a prediction or check-for-understanding.
Explain reporting negative results using a visual and one worked real-world example. Context: Audit a model claim with baselines uncertainty and failure analysis. End with a prediction or check-for-understanding.
Audit a model claim with baselines uncertainty and failure analysis. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Benchmark design, Distribution shift, Robustness and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Combine all three sections: Audit a model claim with baselines uncertainty and failure analysis. 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 Responsible AI, Causality and Scientific Evaluation, 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 →