Week 02 · overview
Week 2: Perceptrons and Layered Networks
A neural network is not a digital brain or a box that “thinks.” It is a parameterized calculation. This week begins with Microsoft's pinned perceptron and multilayer-perceptron chapters, then slows the mechanism down enough to inspect every number.
| Day | Role | Evidence |
|---|---|---|
| 1 | Lesson 1: Perceptron | Trace inputs, weights, bias, score, and threshold |
| 2 | Lesson 2: Multilayer perceptron | Explain why layers and nonlinear activations matter |
| 3 | Investigation | Compare four binary cases and locate a linear boundary |
| 4 | Lab | Build and test a two-layer XOR calculation |
| 5 | Assessment + Reflection | Defend one claim and state the model boundary |
Learning contract
By the end of the week, you will be able to:
- calculate a perceptron output from declared inputs and parameters;
- distinguish training from inference;
- explain why one linear threshold cannot represent XOR;
- trace a small layered network without calling it magic; and
- use a complete four-case test table to support or reject a claim.
Evidence target
Your Layered Network Test Record must preserve all four XOR inputs, both hidden-unit results, the final output, the expected output, and one bounded conclusion. The record proves only that the supplied deterministic calculation matches four supplied cases.
Use no personal or sensitive data. The investigation and Lab are hand-checkable local calculations; they do not train a live model or contact an AI service.
The complete Microsoft source repository—including notebooks, code, images,
assignments, translations, and credits—remains preserved in the local
upstream/ mirror. Review the source record.
Chapter question
How does Perceptrons and Layered Networks change what evidence a student should trust?
What you will understand
You will connect the named mechanism to the evidence it can and cannot support.
What you will do
You will inspect supplied cases, trace the important calculation, and test one bounded claim.
What you will produce
You will produce the week evidence record named in the Lab or reflection.
How the week connects
This week carries the previous AI mechanism forward into a stricter evidence check.
Key vocabulary
mechanism; evidence; boundary; test; claim
Approximate time
Five class meetings at about 210 instructional minutes total.
Final evidence required
Submit the completed week record with at least two exact labels from the artifact.