06. Module 01 Review — Neural Networks Foundations

Focus: init, forward pass, backpropagation, loss functions, mini-batch SGD, activations, optimizers, regularization, scaling-law intuition.

Review loop

1. Skim 02_explainer.md TOC — note any chapter where you are not confident, then re-read just that chapter.
2. Re-answer the self-check questions in 01_weekly_plan.md without looking at notes.
3. Re-do the hardest prompts in 04_daily_recall.md from memory.
4. Sketch the failure-fix table from explainer §6.1 — all 11 rows — without looking.
5. Review 05_hands_on_lab.md and note one implementation bug, one design choice, and one thing you'd improve.
6. Try the hard exercise in explainer §6.5 (numpy XOR MLP) if you have not — confirms you can implement backprop end-to-end.

Reflection

• What clicked this week that previously felt abstract?
• Which part of backprop, init, or optimization still feels fuzzy?
• What should feel automatic before starting Module 02 (Tokenization & Attention)?

Completion gate

• [ ] All 6 explainer chapters read at least once
• [ ] Failure-fix table sketched from memory (all 11 rows)
• [ ] Weekly plan completed
• [ ] MNIST hands_on_lab shipped
• [ ] All daily-recall prompts answerable
• [ ] Ready to move to Module 02