Protocol
Backpropagation Without Autograd
Hand off a NumPy MLP with a numerical-gradient test suite and a bug diary. A reviewer must be able to insert a wrong derivative and see the tests catch it.
40–65 active hours60–90 min sessionsIntermediate
Mastery contract
Core protocol · Standard 1.0 · Hand off a NumPy MLP with a numerical-gradient test suite and a bug diary. A reviewer must be able to insert a wrong derivative and see the tests catch it.
Loading your learning records…
Reasoning — not yet passed
Derive every parameter gradient and explain the chain rule through shared nodes without consulting a solution.
Reliable implementation — not yet passed
Every smooth parameter group passes a relative-error threshold of 1e-5 away from activation kinks.
Reproducible experiment — not yet passed
The model overfits the tiny fixture, while reported learning curves and gradient norms explain failures at bad initial scales.
Defense and handoff — not yet passed
Demonstrate all 3 repaired bugs and defend why a passed gradient check is necessary but not sufficient for generalization.
All four criteria must pass. Activity completion does not satisfy them.
An independent reviewer reruns your work and varies something you did not rehearse: a fresh input, a different working directory, or a declared edge case.
Pilot decisions are provisional, not external credentials. For an appeal or sensitive artifact, contact your designated pilot operator with the submission ID. Export your assessment history.
What you will do
6 total
Derive batched gradients for a two-layer MLP with cross-entropy; annotate matrix shapes and how the batch reduction affects scale.
Implement forward/backward passes and SGD in NumPy without autograd; include a tiny reverse-mode scalar engine to expose graph accumulation.
Overfit a 20-example fixture and compare analytic gradients with finite differences for every parameter group in float64.
Debug an overwritten accumulated gradient, a missing batch divisor and a saturated activation using intermediate gradient logs.
Write a 500–900 word technical report linking the reasoning, code, measurements, and limitations; include the project decision below.
Assess all 4 mastery criteria against saved artifacts, request an independent review, and repeat each failed criterion on a fresh example.
Essentials
A reading session alone never satisfies a mastery criterion.
Store source references, assumptions, and artifact paths beside every result.
Keep an untouched check case that differs from the worked example.
Record all attempts, including failures and results that contradict your prediction.
Use the stated comparison conditions; document every deviation before drawing a conclusion.
Protocol guardrails
Do
- +State the expected result before running the comparison.
- +Keep one minimal reproducible failing case when debugging.
- +Record environment versions and the exact command used.
- +Compare explanations with saved intermediate values.
- +Ask a reviewer to challenge the weakest assumption.
Don't
- ×Do not copy a worked solution and present it as an independent implementation.
- ×Do not tune against held-out evaluation outcomes.
- ×Do not report only the best seed or discard inconvenient runs.
- ×Do not equate elapsed hours or a completed run with a passed assessment.
- ×Do not conceal reduced-scale experiments behind claims about the original full-scale result.
Protocol authorship
Written by NuthinButta as instructional design. The teaching sources are the official references linked in each milestone; the exercises, workload and pass thresholds are ours, not their authors'.