{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:YMDU4HY2SORDUSNGIOVOZKB5XO","short_pith_number":"pith:YMDU4HY2","schema_version":"1.0","canonical_sha256":"c3074e1f1a93a23a49a643aaeca83dbb973a72d49138beae966521259a0414e5","source":{"kind":"arxiv","id":"2606.27216","version":1},"attestation_state":"computed","paper":{"title":"Hierarchical Muon: Tiled Newton-Schulz Updates for Efficient Muon Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.NA"],"primary_cat":"math.NA","authors_text":"Tianshi Xu, Yousef Saad, Yuanzhe Xi, Ziyuan Tang","submitted_at":"2026-06-25T16:05:55Z","abstract_excerpt":"Muon-type optimizers construct update directions for dense neural-network weights by applying a finite Newton-Schulz map to momentum-gradient matrices. For an $H \\times W$ matrix, with $r=\\min\\{H,W\\}$ and $s=\\max\\{H,W\\}$, $K$ steps of the full-matrix Newton-Schulz update require $O(r^2 s K)$ work and couple all rows and columns through repeated Gram matrix products. We introduce Hierarchical Muon (HiMuon), a tiled Newton-Schulz scheme for Muon-type optimization. HiMuon partitions each momentum-gradient matrix into $T \\times T$ tiles, applies the same finite Newton-Schulz map independently to e"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.27216","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.NA","submitted_at":"2026-06-25T16:05:55Z","cross_cats_sorted":["cs.LG","cs.NA"],"title_canon_sha256":"2aee6b6294d4073841c6bf9a950c144a53659e611e56895058dbe8ea48a600c5","abstract_canon_sha256":"1b5cb276b949b490f101fd2d848d86c6c9dba789500e594bbe7dbf5a1a0155eb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-26T01:16:13.810810Z","signature_b64":"Pwv5HmyBuhc+fDRgLIhAS1vcY59wZjQ+Epq5cdpNUvXtyGHIT3zptZTxHoGyglX77YdRnH1H/CGoRViak0dMDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c3074e1f1a93a23a49a643aaeca83dbb973a72d49138beae966521259a0414e5","last_reissued_at":"2026-06-26T01:16:13.810401Z","signature_status":"signed_v1","first_computed_at":"2026-06-26T01:16:13.810401Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hierarchical Muon: Tiled Newton-Schulz Updates for Efficient Muon Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.NA"],"primary_cat":"math.NA","authors_text":"Tianshi Xu, Yousef Saad, Yuanzhe Xi, Ziyuan Tang","submitted_at":"2026-06-25T16:05:55Z","abstract_excerpt":"Muon-type optimizers construct update directions for dense neural-network weights by applying a finite Newton-Schulz map to momentum-gradient matrices. For an $H \\times W$ matrix, with $r=\\min\\{H,W\\}$ and $s=\\max\\{H,W\\}$, $K$ steps of the full-matrix Newton-Schulz update require $O(r^2 s K)$ work and couple all rows and columns through repeated Gram matrix products. We introduce Hierarchical Muon (HiMuon), a tiled Newton-Schulz scheme for Muon-type optimization. HiMuon partitions each momentum-gradient matrix into $T \\times T$ tiles, applies the same finite Newton-Schulz map independently to e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.27216","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.27216/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.27216","created_at":"2026-06-26T01:16:13.810467+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.27216v1","created_at":"2026-06-26T01:16:13.810467+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.27216","created_at":"2026-06-26T01:16:13.810467+00:00"},{"alias_kind":"pith_short_12","alias_value":"YMDU4HY2SORD","created_at":"2026-06-26T01:16:13.810467+00:00"},{"alias_kind":"pith_short_16","alias_value":"YMDU4HY2SORDUSNG","created_at":"2026-06-26T01:16:13.810467+00:00"},{"alias_kind":"pith_short_8","alias_value":"YMDU4HY2","created_at":"2026-06-26T01:16:13.810467+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO","json":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO.json","graph_json":"https://pith.science/api/pith-number/YMDU4HY2SORDUSNGIOVOZKB5XO/graph.json","events_json":"https://pith.science/api/pith-number/YMDU4HY2SORDUSNGIOVOZKB5XO/events.json","paper":"https://pith.science/paper/YMDU4HY2"},"agent_actions":{"view_html":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO","download_json":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO.json","view_paper":"https://pith.science/paper/YMDU4HY2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.27216&json=true","fetch_graph":"https://pith.science/api/pith-number/YMDU4HY2SORDUSNGIOVOZKB5XO/graph.json","fetch_events":"https://pith.science/api/pith-number/YMDU4HY2SORDUSNGIOVOZKB5XO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO/action/storage_attestation","attest_author":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO/action/author_attestation","sign_citation":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO/action/citation_signature","submit_replication":"https://pith.science/pith/YMDU4HY2SORDUSNGIOVOZKB5XO/action/replication_record"}},"created_at":"2026-06-26T01:16:13.810467+00:00","updated_at":"2026-06-26T01:16:13.810467+00:00"}