◄ WORLD II · THE FOLDTHE OCHO · blue builds │ the machine │ red breaks

THE DENOISING AUTOENCODER

Corrupt the data, then learn to un-corrupt it. Vincent's 2008 move looks like a trick and turns out to be a theorem: the best reconstructor of a noised point is the posterior mean E[x | x̃], and its correction vector r*(x̃) − x̃ is exactly σ² ∇log p(x̃) — it points back up the density, toward where the data lives. Down the center: clean points go in, get noised, the optimal map pulls them out onto the manifold. The blue team proves the map is the score; the red team trains it at zero noise so it learns nothing.

source Vincent, Larochelle, Bengio & Manzagol, Extracting and Composing Robust Features with Denoising Autoencoders, ICML 2008, 1096–1103 — doi:10.1145/1390156.1390294 (no stable arXiv id; DOI cited, marked AMBER). The score identity is Vincent 2011 & Alain–Bengio 2014. Rendered, not quoted.

◧ blue team · builds & defends
3

THE MODEL — the posterior mean

Data is Gaussian, x ∼ N(μ, σd²); corruption adds independent noise, x̃ = x + ε, ε ∼ N(0, σn²). The reconstructor that minimizes E∥r(x̃)−x∥² is not learned by luck — it is the conditional mean, and for Gaussians it is closed-form:

r*(x̃) = a x̃ + (1−a)μ,   a = σd² / (σd²+σn²)

A precision-weighted average of the noisy point and the prior mean. The map is a contraction toward the data (slope a < 1), and its Bayes risk is exactly a σn² per coordinate.

5

THE LINEAGE — the score in disguise AVAN

Subtract the input from the reconstruction and something familiar appears:

r*(x̃) − x̃ = σn² ∇log p(x̃)

The correction is the score of the corrupted density, scaled by the noise variance — Vincent (2011), Alain–Bengio (2014). A denoiser is a score estimator. Scale it across many noise levels and you have the-score-matching, which the-diffusion-model ran with. Each sphere is the next one's premise.

7

THE WITNESS live

The blue team's live check: re-derive the correction field from the current controls and confirm r*−x̃ = σn²∇log p to machine precision at sample points. If red trains at zero noise, the field collapses and this badge turns red.

▼ the machine ▼
4

DATA IN — clean & corrupted in ↓

Two isotropic Gaussians in the plane, mean at the origin. Clean points are drawn from the data density N(0,σd²I); each is then kicked by fresh noise N(0,σn²I) to give a corrupted point off the manifold. Every sample uses a fixed-seed PRNG, so the picture is deterministic and the self-check reproducible.

The task the machine is handed: given only a corrupted point, name the clean one it most likely came from.

▼   feed the corrupted cloud into the engine   ▼
0

▣ THE PANEL — the reconstruction field LIT

Pink arrows are the correction r*(x̃)−x̃ at grid points — every one points inward, up the density, toward the clean cloud. Drop σn to 0 and the arrows vanish: on the manifold the optimal map is the identity.

Every number is computed live from the two σ values — contraction a, Bayes risk n², and the max mismatch between the correction and σn²∇log p across the grid. Nothing is looked up.

▼   the engine emits denoised points + a proof   ▼
8

DATA OUT — the result out ↓

What the machine produces, proven: the optimal denoiser is the posterior mean (checked against numeric integration to <1e−4), its correction equals the score exactly (<1e−9), it is the identity at zero noise, it strictly beats the do-nothing map in Monte-Carlo MSE, and larger corruption forces a smoother (more contractive) map that keeps only coarse structure.

The blue team's witness (left) re-derives the score identity live; the red team (right) tries to make the field lie.

red team · attacks & breaks ◨
1

THE ADVERSARY

WALL The clean closed form is a Gaussian conceit. Real data is not one Gaussian, and then E[x|x̃] has no formula — it is a hard non-linear map you must actually learn, and the training network can under-fit it, memorize, or collapse to the mean.

The score identity is exact only in expectation and cleanest for small σ; a single fixed noise level sees the density through one blur and misses fine structure — the reason diffusion needs a ladder of noise scales, not one denoiser. And a denoiser trained on the wrong noise level is confidently wrong.

2

THE GRAVEYARD

"An autoencoder just learns to copy its input." Cut. A plain autoencoder can; a denoising one cannot — the corruption forbids the identity and forces it to model where the data lives.

"The denoiser removes the noise it was given." Corrected. It cannot recover the specific ε; it returns the mean of all clean points that could have produced x̃ — the manifold projection, not the inverse of one draw.

"Bigger corruption is strictly better regularization." Cut. Bigger σn smooths more (a→0, map→constant mean); past a point it erases the very structure you wanted — a bias–variance knob, not a free lunch.

6

THE TAMPER — break it

The red team's move: train the denoiser at zero corruption. Then its optimal map is the plain identity — it reconstructs perfectly on clean data and learns nothing about the manifold: r−x̃ = 0, no score, no inward field.

Set the training noise to zero and the arrows collapse to nothing. The witness (window 7) re-derives the score identity, finds r−x̃ = 0 ≠ σn²∇log p, and turns red. The attack is real and it is caught.