{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:PCUROFJ3NP2BXPM6BK2AMSCFQV","short_pith_number":"pith:PCUROFJ3","schema_version":"1.0","canonical_sha256":"78a917153b6bf41bbd9e0ab4064845855d063da1352f286aef200709c7c6d198","source":{"kind":"arxiv","id":"2605.16165","version":1},"attestation_state":"computed","paper":{"title":"Second-Order Multi-Level Variance Correction for Modality Competition in Multimodal Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Wes Armour, Yishun Lu","submitted_at":"2026-05-15T16:45:56Z","abstract_excerpt":"Autoregressive next-token training offers a unified formulation for image generation and text understanding, but it also creates strong modality competition that destabilizes optimization and limits large-batch scaling. We show that first-order optimizers such as AdamW are vulnerable to cross-modality gradient heterogeneity, while second-order preconditioning, particularly SOAP, provides a more stable basis for multimodal alignment. Building on this insight, we propose \\emph{ML-FOP-SOAP}, a second-order optimization framework with Multi-Level Variance Correction. Our Fisher-Orthogonal Projecti"},"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":"2605.16165","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-15T16:45:56Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a3f0d08d59584d1d45999f076501f46bbf2b9074d9a494d614601bf5ccfb85d1","abstract_canon_sha256":"048bf64fb3cae2100c666ac2543e088f5fb77258e21dfceb7c6df187d8ab69a9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:55.740333Z","signature_b64":"jV/jNK4Jcu50C4j8KW5XkmNrKqzBpas93v/Qqd3DmrSVO90QFwLu5mD6fhcT+D4G0ubqahFGON7Nvd8likbCAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"78a917153b6bf41bbd9e0ab4064845855d063da1352f286aef200709c7c6d198","last_reissued_at":"2026-05-20T00:01:55.739517Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:55.739517Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Second-Order Multi-Level Variance Correction for Modality Competition in Multimodal Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Wes Armour, Yishun Lu","submitted_at":"2026-05-15T16:45:56Z","abstract_excerpt":"Autoregressive next-token training offers a unified formulation for image generation and text understanding, but it also creates strong modality competition that destabilizes optimization and limits large-batch scaling. We show that first-order optimizers such as AdamW are vulnerable to cross-modality gradient heterogeneity, while second-order preconditioning, particularly SOAP, provides a more stable basis for multimodal alignment. Building on this insight, we propose \\emph{ML-FOP-SOAP}, a second-order optimization framework with Multi-Level Variance Correction. Our Fisher-Orthogonal Projecti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.16165","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/2605.16165/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"cited_work_retraction","ran_at":"2026-05-19T18:21:56.372264Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:31.034385Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"external_links","ran_at":"2026-05-19T17:31:47.003238Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T16:41:55.433531Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"4db0a6308319809513b6898e8bfadcada9af9d88477a7e08bd265669a7bf624b"},"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":"2605.16165","created_at":"2026-05-20T00:01:55.739637+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.16165v1","created_at":"2026-05-20T00:01:55.739637+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16165","created_at":"2026-05-20T00:01:55.739637+00:00"},{"alias_kind":"pith_short_12","alias_value":"PCUROFJ3NP2B","created_at":"2026-05-20T00:01:55.739637+00:00"},{"alias_kind":"pith_short_16","alias_value":"PCUROFJ3NP2BXPM6","created_at":"2026-05-20T00:01:55.739637+00:00"},{"alias_kind":"pith_short_8","alias_value":"PCUROFJ3","created_at":"2026-05-20T00:01:55.739637+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/PCUROFJ3NP2BXPM6BK2AMSCFQV","json":"https://pith.science/pith/PCUROFJ3NP2BXPM6BK2AMSCFQV.json","graph_json":"https://pith.science/api/pith-number/PCUROFJ3NP2BXPM6BK2AMSCFQV/graph.json","events_json":"https://pith.science/api/pith-number/PCUROFJ3NP2BXPM6BK2AMSCFQV/events.json","paper":"https://pith.science/paper/PCUROFJ3"},"agent_actions":{"view_html":"https://pith.science/pith/PCUROFJ3NP2BXPM6BK2AMSCFQV","download_json":"https://pith.science/pith/PCUROFJ3NP2BXPM6BK2AMSCFQV.json","view_paper":"https://pith.science/paper/PCUROFJ3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.16165&json=true","fetch_graph":"https://pith.science/api/pith-number/PCUROFJ3NP2BXPM6BK2AMSCFQV/graph.json","fetch_events":"https://pith.science/api/pith-number/PCUROFJ3NP2BXPM6BK2AMSCFQV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PCUROFJ3NP2BXPM6BK2AMSCFQV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PCUROFJ3NP2BXPM6BK2AMSCFQV/action/storage_attestation","attest_author":"https://pith.science/pith/PCUROFJ3NP2BXPM6BK2AMSCFQV/action/author_attestation","sign_citation":"https://pith.science/pith/PCUROFJ3NP2BXPM6BK2AMSCFQV/action/citation_signature","submit_replication":"https://pith.science/pith/PCUROFJ3NP2BXPM6BK2AMSCFQV/action/replication_record"}},"created_at":"2026-05-20T00:01:55.739637+00:00","updated_at":"2026-05-20T00:01:55.739637+00:00"}