pith. sign in
Pith Number

pith:GAIYONA2

pith:2026:GAIYONA27QJZ5WZRI3A54XH5KW
not attested not anchored not stored refs pending

DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices

Chaojun Xiao, Chenyang Song, Weilin Zhao, Xu Han, Yingfa Chen, Zhiyuan Liu

DECO sparse MoE matches dense Transformer performance while activating only 20% of experts.

arxiv:2605.10933 v3 · 2026-05-11 · cs.LG · cs.CL

Add to your LaTeX paper
\usepackage{pith}
\pithnumber{GAIYONA27QJZ5WZRI3A54XH5KW}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp
2 Internet Archive
3 Author claim open · sign in to claim
4 Citations open
5 Replications open
Portable graph bundle live · download bundle · merged state
The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

DECO, activating only 20% of experts, matches dense performance and outperforms established MoE baselines.

C2weakest assumption

That the learned expert-wise scaling and NormSiLU will continue to produce stable sparsity and matching performance when model size, data distribution, or hardware change substantially.

C3one line summary

DECO matches dense model performance at 20% expert activation via ReLU-based routing with learnable scaling and the NormSiLU activation, plus a 3x real-hardware speedup.

Formal links

2 machine-checked theorem links

Receipt and verification
First computed 2026-05-21T02:05:05.288209Z
Builder pith-number-builder-2026-05-17-v1
Signature Pith Ed25519 (pith-v1-2026-05) · public key
Schema pith-number/v1.0

Canonical hash

301187341afc139edb3146c1de5cfd559fabbcbc7617d1fb69e5b6e30d1b9e42

Aliases

arxiv: 2605.10933 · arxiv_version: 2605.10933v3 · doi: 10.48550/arxiv.2605.10933 · pith_short_12: GAIYONA27QJZ · pith_short_16: GAIYONA27QJZ5WZR · pith_short_8: GAIYONA2
Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/GAIYONA27QJZ5WZRI3A54XH5KW \
  | 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: 301187341afc139edb3146c1de5cfd559fabbcbc7617d1fb69e5b6e30d1b9e42
Canonical record JSON
{
  "metadata": {
    "abstract_canon_sha256": "132fa94163a8ca5817b3b05d10ac9448834d853755a9e12f470929fab39dc91a",
    "cross_cats_sorted": [
      "cs.CL"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-11T17:58:28Z",
    "title_canon_sha256": "f02b018379c25ddd549c16fb7d757a3ba9e8fb0b63074a6bad757b66cc6e9974"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.10933",
    "kind": "arxiv",
    "version": 3
  }
}