Level-3 seal for the trained-CNN witness #
The MLP rung's pdiv_fwd/trainedMlp_backward_nontrivial program at the
CONVOLUTIONAL witness: one whole-net Jacobian entry of the trained CNN,
computed in closed form by peeling mnistCnnNoBnForward with pdiv_comp
from the output side. Exact backward-cotangent tables (all in-kernel
rationals): dense head slices (t4V/t3V), the max-pool argmax routing
(t2V, via MaxPool2IsArgmax at each of the 72 positions), the ReLU mask
folds (m2V/m1V), and the conv input-VJPs (t1V and the final entry,
via conv2d_input_grad_formula through HasVJPAt.correct).
The sealed entry: ∂ logit_7 / ∂ pixel (0,2) at the witness =
-326103939411/34359738368 ≈ -9.4909 ≠ 0, hence
trainedCnn_backward_nontrivial (the proven backward is not the zero map),
trainedCnn_jacobian_nonzero (fderiv ≠ 0), and trainedCnn_not_constant.
Generated by scripts/trained_cnn_seal.py; tables are DATA here.
- relu∘dense3:
pdiv (head) p2f · 7.
Equations
- One or more equations did not get rendered due to their size.
Instances For
relu mask₂ ⊙ t2V, tensor form (the conv2 backward cotangent).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Instances For
relu mask₁ ⊙ t1V, tensor form (the conv1 backward cotangent).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Instances For
conv2 at the relu(conv1) table = the conv2 table, flat form.
conv1 at the witness input = the conv1 table, flat form.
The conv2 layer witness, type-ascribed at flatConv (defeq).
Equations
Instances For
The conv1 layer witness, type-ascribed at flatConv (defeq).
Equations
Instances For
Level 3: the trained-weight CNN backward is not the zero map — the seal the MLP rung carries, now at the convolutional witness.