Protocol
Variational Autoencoders and Latent Evidence
Produce a VAE experiment with a collapse diagnosis and a defended likelihood model. Preserve unsuccessful runs and separate density claims from sample aesthetics.
40–70 active hours60–90 min sessionsAdvanced
Mastery contract
Specialization · Standard 1.0 · Produce a VAE experiment with a collapse diagnosis and a defended likelihood model. Preserve unsuccessful runs and separate density claims from sample aesthetics.
Loading your learning records…
Reasoning — not yet passed
Derive the lower bound and identify when its gap equals a posterior KL divergence.
Reliable implementation — not yet passed
The loss uses a likelihood consistent with the data and passes KL/value checks on a known Gaussian fixture.
Reproducible experiment — not yet passed
Compare the declared reference setting and reconstruction/regularization settings on identical held-out data and budgets; retain seeds, objective components and uncertainty, not only samples.
Defense and handoff — not yet passed
Use measured reconstruction/KL results to defend the tradeoff, diagnose one collapse case and explain with a counterexample why a latent coordinate is not automatically causal.
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 the ELBO and diagonal-Gaussian KL; explain which expectation is differentiated through the reparameterization.
Implement a small VAE in PyTorch with separate reconstruction and KL logs; verify the analytic KL against a Monte Carlo estimate.
Compare 3 latent sizes or KL weights over 3 seeds; report both objective components, reconstructions and samples under matched data splits.
Debug an incorrect variance parameterization and posterior collapse; compare a decoder-capacity intervention with a KL-weight intervention.
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'.