pith:5OAF7MUB
Unlearning with Asymmetric Sources: Improved Unlearning-Utility Trade-off with Public Data
Asymmetric Langevin Unlearning injects public data to reduce certified unlearning noise by a factor of O(1/n_pub²) while preserving model utility.
arxiv:2605.11170 v2 · 2026-05-11 · cs.LG · cs.CR
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{5OAF7MUBU76G5JQ24MI4BDQE35}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
We prove that public data injection suppresses the unlearning cost by a factor of O(1/n_pub²), guaranteeing a strict computational advantage over retraining.
The analysis assumes that the underlying Langevin dynamics and variational Rényi divergence bounds remain valid when public data is injected asymmetrically, and that the distribution mismatch between public and private sources can be explicitly characterized without invalidating the utility guarantees.
Asymmetric Langevin Unlearning uses public data to suppress unlearning noise costs by O(1/n_pub²), enabling practical mass unlearning with preserved utility under distribution mismatch.
Formal links
Receipt and verification
| First computed | 2026-06-03T01:05:51.307619Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
eb805fb281a7fc6ea61ae311c08e04df5a574ac979d21a632ddc90fcc66282ef
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/5OAF7MUBU76G5JQ24MI4BDQE35 \
| 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: eb805fb281a7fc6ea61ae311c08e04df5a574ac979d21a632ddc90fcc66282ef
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "b7cc1be403590cc7cf4ec48b027feabd34da356b9574194cb1b061b65ef6aa67",
"cross_cats_sorted": [
"cs.CR"
],
"license": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
"primary_cat": "cs.LG",
"submitted_at": "2026-05-11T19:28:33Z",
"title_canon_sha256": "417dd600062dd000d608075e6450b915803bd4b36ca57bfd2b0712414b99c7bf"
},
"schema_version": "1.0",
"source": {
"id": "2605.11170",
"kind": "arxiv",
"version": 2
}
}