pith:H47YDJN6
ZenBrain: A Neuroscience-Inspired 7-Layer Memory Architecture for Autonomous AI Systems
A 7-layer neuroscience-inspired memory system for AI reaches 91 percent of long-context oracle accuracy at 1/106th the token cost.
arxiv:2604.23878 v3 · 2026-04-26 · cs.AI · cs.LG
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{H47YDJN6QLTUEOMDTYUCMZAPG6}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
On LongMemEval-500, ZenBrain matches a long-context oracle's binary-judge accuracy to within 4.5 pp (47.7% vs. 52.2%; 91.3%) at 1/106th of the per-query token cost, and wins all 12 head-to-head answer-quality cells against Letta, Mem0, and A-Mem under Bonferroni correction.
That the 15 neuroscience mechanisms translate directly into effective AI components without introducing hidden interactions or benchmark-specific artifacts, and that the 60-day stress ablations with 10 seeds fully isolate each mechanism's contribution.
ZenBrain unifies 15 neuroscience mechanisms into a 7-layer memory system that achieves near-oracle long-context accuracy at 1/106th token cost and outperforms prior memory architectures in controlled comparisons.
Cited by
Receipt and verification
| First computed | 2026-08-11T01:20:42.478004Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
3f3f81a5be82e74239839e2826640f37b9ef78fcb07785b0d2ae2fea5bfe1685
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/H47YDJN6QLTUEOMDTYUCMZAPG6 \
| 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: 3f3f81a5be82e74239839e2826640f37b9ef78fcb07785b0d2ae2fea5bfe1685
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "64a15dd32a35ce611e8255ddf4f18a94360c851b3b45b0f8c5e09b80b1997692",
"cross_cats_sorted": [
"cs.LG"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.AI",
"submitted_at": "2026-04-26T20:39:19Z",
"title_canon_sha256": "c2f94e3057661e312d8f27ecbf417ec5149b318e4c2d9ce5d9c91890b750c5fd"
},
"schema_version": "1.0",
"source": {
"id": "2604.23878",
"kind": "arxiv",
"version": 3
}
}