Ship and explain the single-neuron trainer

A shippable lab artifact includes: how to run it, the metric result (final predictions match gate table; loss decreases across epochs), the failure you can still reproduce (learning rate so large weights diverge, or claiming XOR success with one neuron), the security gate for hard-coding 'trained' weights without a verifiable training loop, and a rollback note. The user decision it supports remains: learn weights for a linearly separable gate by gradient steps you can print.

cases=[((0,0),0.13),((0,1),0.91),((1,0),0.91),((1,1),1.0)]
assert [int(p>=.5) for _,p in cases]==[0,1,1,1]
print('OR gate shipped',cases)

Expected evidence: stage evidence for mastery ship. Store this beside the fixture version so scores remain meaningful after content changes in neural-nets-by-building.

State operating limits for the single-neuron trainer in plain language: fixture size, offline vs live dependencies, and what would require a new eval set. Shipping neural-nets-by-building is honest scoping, not maximal confidence language.

Fill this table in your notes for the single-neuron trainer:

• Proved on AND gate truth table: …
• Unproved beyond the fixture: …
• Metric that blocks release: final predictions match gate table; loss decreases across epochs
• Failure still reproducible: learning rate so large weights diverge, or claiming XOR success with one neuron
• Security gate: hard-coding 'trained' weights without a verifiable training loop
• Rollback: …

Ship the narrative only when the unproved list is honest. Reviewers trust narrow claims that support learn weights for a linearly separable gate by gradient steps you can print more than maximal language that collapses under the first production oddity.

Hand your ship note to a peer and ask them to recreate a proved/unproved ship note with rollback without watching you type. If they cannot, your evidence is still tribal knowledge. Tighten the run command and the metric line until a stranger can validate the single-neuron trainer against AND gate truth table.

At the mastery and shipping stage for neural-nets-by-building, the job is narrower than finishing a product demo. You are creating one progressive evidence piece about AND gate truth table that later pages inherit without redefining success. Keep that fixture small enough to inspect by hand, keep outputs copy-pasteable as text, and refuse to narrate this baseline as if it were a production SLA: untrained weight predictions on the gate table.

For this page specifically, success looks like a proved/unproved ship note with rollback while still centering the user decision to learn weights for a linearly separable gate by gradient steps you can print. If you cannot point to a file, command, or assertion that proves that for the single-neuron trainer, stay on this page instead of advancing.

Glossary: deep learning · Glossary: gradient descent

Previous