pith:3VKFXRQM
A Theory of Saddle Escape in Deep Nonlinear Networks
An exact identity on Frobenius norm imbalances in layer weights reduces deep nonlinear training to a scalar ODE whose escape time scales as ε to the power of minus (r minus 2), where r is the number of bottleneck layers.
arxiv:2605.01288 v3 · 2026-05-02 · cs.LG · cond-mat.dis-nn · stat.ML
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{3VKFXRQM6RXARY7J43WUMGYQRO}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
We derive an exact identity for the imbalance of Frobenius norms of layer weight matrices that holds for any smooth activation and any differentiable loss [...] giving a critical-depth escape time law τ★ = Θ(ε^{-(r-2)}) governed by the number r of layers at the bottleneck scale rather than the total depth L.
The reduction to a scalar ODE relies on an approximate balance law on the permutation-symmetric submanifold whose accuracy and range of validity are not derived from first principles but stated as holding approximately.
An exact norm-imbalance identity classifies activations into four classes and reduces deep nonlinear training flow to a scalar ODE that predicts saddle escape time scaling as ε to the power of minus (r-2) for r bottleneck layers.
Formal links
Receipt and verification
| First computed | 2026-06-24T01:15:03.268859Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
dd545bc60cf46e08e3e9e6ed461b108b829581fa3e9cacc5ccac9f69c012f4bf
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/3VKFXRQM6RXARY7J43WUMGYQRO \
| jq -c '.canonical_record' \
| python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: dd545bc60cf46e08e3e9e6ed461b108b829581fa3e9cacc5ccac9f69c012f4bf
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "8635d78054aeba55bcc86b93793aeb1cb6e41a0c5a13564e20ba2274e366a0c6",
"cross_cats_sorted": [
"cond-mat.dis-nn",
"stat.ML"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.LG",
"submitted_at": "2026-05-02T06:55:15Z",
"title_canon_sha256": "52163603e99629cbdd0bb6cb17146be35167df9f9963c970d5f3e49cf163a903"
},
"schema_version": "1.0",
"source": {
"id": "2605.01288",
"kind": "arxiv",
"version": 3
}
}