AI-06 · 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 | Deep Learning Foundations
Playlist title reserved in the plan; no YouTube playlist or upload has been created yet.
Sign in to save this learning pathPractical outcome: Trace a two-layer network with a tiny hand-checkable batch.
Explain artificial neuron using a visual and one worked real-world example. Context: Trace a two-layer network with a tiny hand-checkable batch. End with a prediction or check-for-understanding.
Explain activations using a visual and one worked real-world example. Context: Trace a two-layer network with a tiny hand-checkable batch. End with a prediction or check-for-understanding.
Explain layer shapes using a visual and one worked real-world example. Context: Trace a two-layer network with a tiny hand-checkable batch. End with a prediction or check-for-understanding.
Explain forward pass using a visual and one worked real-world example. Context: Trace a two-layer network with a tiny hand-checkable batch. End with a prediction or check-for-understanding.
Use the first concepts from this section. Trace a two-layer network with a tiny hand-checkable batch. 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 loss functions using a visual and one worked real-world example. Context: Trace a two-layer network with a tiny hand-checkable batch. End with a prediction or check-for-understanding.
Explain backpropagation using a visual and one worked real-world example. Context: Trace a two-layer network with a tiny hand-checkable batch. End with a prediction or check-for-understanding.
Explain computational graphs using a visual and one worked real-world example. Context: Trace a two-layer network with a tiny hand-checkable batch. End with a prediction or check-for-understanding.
Explain automatic differentiation using a visual and one worked real-world example. Context: Trace a two-layer network with a tiny hand-checkable batch. End with a prediction or check-for-understanding.
Trace a two-layer network with a tiny hand-checkable batch. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Artificial neuron, Activations, Layer shapes and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Practical outcome: Diagnose unstable training using supplied curves.
Explain weight initialisation using a visual and one worked real-world example. Context: Diagnose unstable training using supplied curves. End with a prediction or check-for-understanding.
Explain mini-batches using a visual and one worked real-world example. Context: Diagnose unstable training using supplied curves. End with a prediction or check-for-understanding.
Explain learning rates using a visual and one worked real-world example. Context: Diagnose unstable training using supplied curves. End with a prediction or check-for-understanding.
Explain momentum using a visual and one worked real-world example. Context: Diagnose unstable training using supplied curves. End with a prediction or check-for-understanding.
Use the first concepts from this section. Diagnose unstable training using supplied curves. 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 adaptive optimisers using a visual and one worked real-world example. Context: Diagnose unstable training using supplied curves. End with a prediction or check-for-understanding.
Explain vanishing gradients using a visual and one worked real-world example. Context: Diagnose unstable training using supplied curves. End with a prediction or check-for-understanding.
Explain exploding gradients using a visual and one worked real-world example. Context: Diagnose unstable training using supplied curves. End with a prediction or check-for-understanding.
Explain gradient clipping using a visual and one worked real-world example. Context: Diagnose unstable training using supplied curves. End with a prediction or check-for-understanding.
Diagnose unstable training using supplied curves. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Weight initialisation, Mini-batches, Learning rates and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Practical outcome: Compare model changes while holding data and evaluation fixed.
Explain regularisation using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Explain dropout using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Explain normalisation using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Explain early stopping using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Use the first concepts from this section. Compare model changes while holding data and evaluation fixed. 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 data augmentation using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Explain learning curves using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Explain ablation studies using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Explain reproducible seeds using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Explain compute budgets using a visual and one worked real-world example. Context: Compare model changes while holding data and evaluation fixed. End with a prediction or check-for-understanding.
Compare model changes while holding data and evaluation fixed. Work without the example, compare two cases, record assumptions and limitations, then self-review correctness, reasoning and clarity.
6–12 questions covering Regularisation, Dropout, Normalisation and the remaining section concepts, with cumulative retrieval from earlier sections. Include a new transfer problem; 70% unlocks progression.
Combine all three sections: Compare model changes while holding data and evaluation fixed. 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 Deep Learning Foundations, 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 →