pith:6WJTXNHU
Most ReLU Networks Admit Identifiable Parameters
ReLU networks with input and hidden widths at least two admit an open set of identifiable parameters.
arxiv:2605.03601 v2 · 2026-05-05 · cs.LG · cs.DM · math.CO
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{6WJTXNHUBJJOKPN55B42GDIQC7}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
Our main result shows that for every architecture whose input and hidden layers have width at least two, there exists an open set of identifiable parameters. This implies that the functional dimension of every such architecture is exactly the number of parameters minus the number of hidden neurons.
The assumption that the architecture has input and hidden layer widths at least two; the result is stated only for open sets of parameters and does not claim identifiability for every parameter vector.
For ReLU networks with input and hidden widths at least 2, most parameters are identifiable up to symmetry, so the functional dimension equals the parameter count minus the number of hidden neurons.
Receipt and verification
| First computed | 2026-05-21T01:04:26.728429Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
f5933bb4f40a52e53dbde879a30d1017c7edc2e69231f152d426bace7d206a21
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/6WJTXNHUBJJOKPN55B42GDIQC7 \
| 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: f5933bb4f40a52e53dbde879a30d1017c7edc2e69231f152d426bace7d206a21
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "e1b1a8bd998b8122674b64d1b2627b4781114aa017e7d3ea2b12b63cc0042308",
"cross_cats_sorted": [
"cs.DM",
"math.CO"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.LG",
"submitted_at": "2026-05-05T10:25:28Z",
"title_canon_sha256": "8d8de95237a580c8e2cc2a2929b8bfbaf69d3c1624b9d6c310908c60cdca5d60"
},
"schema_version": "1.0",
"source": {
"id": "2605.03601",
"kind": "arxiv",
"version": 2
}
}