{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GAIYONA27QJZ5WZRI3A54XH5KW","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"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}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.10933","created_at":"2026-05-21T02:05:05Z"},{"alias_kind":"arxiv_version","alias_value":"2605.10933v3","created_at":"2026-05-21T02:05:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.10933","created_at":"2026-05-21T02:05:05Z"},{"alias_kind":"pith_short_12","alias_value":"GAIYONA27QJZ","created_at":"2026-05-21T02:05:05Z"},{"alias_kind":"pith_short_16","alias_value":"GAIYONA27QJZ5WZR","created_at":"2026-05-21T02:05:05Z"},{"alias_kind":"pith_short_8","alias_value":"GAIYONA2","created_at":"2026-05-21T02:05:05Z"}],"graph_snapshots":[{"event_id":"sha256:a666e985f15ebfbde823bb74481014b0171771c45898ed5ec8d648f4137cb33b","target":"graph","created_at":"2026-05-21T02:05:05Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"DECO, activating only 20% of experts, matches dense performance and outperforms established MoE baselines."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"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."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"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."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"DECO sparse MoE matches dense Transformer performance while activating only 20% of experts."}],"snapshot_sha256":"fb0d46a45ae445bdec59bbd573a2ac45d8e9b9807b9a5e97c2abb680d0d9a919"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b3f92bf15c7f14242316b7185ec75741ceb26ba35c9d21eee0ef878cf7d36495"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-20T05:02:00.913621Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T13:36:55.079203Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T10:31:17.132399Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T08:53:45.739118Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.10933/integrity.json","findings":[],"snapshot_sha256":"96fadb87ebbca95e4bdf70d6c14a464c9c1764e752eea7a100432d28aacfb521","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"While Mixture-of-Experts (MoE) scales model capacity without proportionally increasing computation, its massive total parameter footprint creates significant storage and memory-access bottlenecks, which hinder efficient end-side deployment that simultaneously requires high performance, low computational cost, and small storage overhead. To achieve these properties, we present DECO, a sparse MoE architecture designed to match the performance of dense Transformers under identical total parameter budgets and training tokens. DECO utilizes the differentiable and flexible ReLU-based routing enhance","authors_text":"Chaojun Xiao, Chenyang Song, Weilin Zhao, Xu Han, Yingfa Chen, Zhiyuan Liu","cross_cats":["cs.CL"],"headline":"DECO sparse MoE matches dense Transformer performance while activating only 20% of experts.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-11T17:58:28Z","title":"DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.10933","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-13T07:25:41.711109Z","id":"b57856c1-ac6a-44b7-9016-87aeeeded385","model_set":{"reader":"grok-4.3"},"one_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.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"DECO sparse MoE matches dense Transformer performance while activating only 20% of experts.","strongest_claim":"DECO, activating only 20% of experts, matches dense performance and outperforms established MoE baselines.","weakest_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."}},"verdict_id":"b57856c1-ac6a-44b7-9016-87aeeeded385"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c2760bc455e601a37a0d943786af64c9af788e72bad45df06948d4d428b675a2","target":"record","created_at":"2026-05-21T02:05:05Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"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}},"canonical_sha256":"301187341afc139edb3146c1de5cfd559fabbcbc7617d1fb69e5b6e30d1b9e42","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"301187341afc139edb3146c1de5cfd559fabbcbc7617d1fb69e5b6e30d1b9e42","first_computed_at":"2026-05-21T02:05:05.288209Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-21T02:05:05.288209Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KfnrNMTwO0PUNkImjYosp5QCOvoyaLrMhkWbOoAQFECdep9edp9s42QHHjePm1+sefpTJZm5emZE2Jse6MbsDA==","signature_status":"signed_v1","signed_at":"2026-05-21T02:05:05.289148Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.10933","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c2760bc455e601a37a0d943786af64c9af788e72bad45df06948d4d428b675a2","sha256:a666e985f15ebfbde823bb74481014b0171771c45898ed5ec8d648f4137cb33b"],"state_sha256":"c383eefcb37ba1b4a80d1071229516beb45ea29ccf02d522173e91e5bfce0591"}