{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2S6NTSNWQHHKBUUHZG4Z4DTZIH","short_pith_number":"pith:2S6NTSNW","schema_version":"1.0","canonical_sha256":"d4bcd9c9b681cea0d287c9b99e0e7941c44117f5d574d3b87a011793317aa04e","source":{"kind":"arxiv","id":"2605.01046","version":3},"attestation_state":"computed","paper":{"title":"Learning in the Fisher Subspace: A Guided Initialization for LoRA Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hung-Yu Kao, Ying-Jia Lin, Zhi-Quan Feng","submitted_at":"2026-05-01T19:20:25Z","abstract_excerpt":"LoRA adapts large language models (LLMs) by restricting updates to low-rank subspaces of pre-trained weights. While this substantially reduces training cost, the effectiveness of adaptation critically depends on which subspace is chosen at initialization: a poor initialization that allocates capacity to task-irrelevant directions can severely hinder downstream performance. Existing initialization strategies primarily rely on the intrinsic properties of pre-trained weights, implicitly assuming that weight geometry alone reflects task relevance. However, such criteria overlook how the model inte"},"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":"2605.01046","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-01T19:20:25Z","cross_cats_sorted":[],"title_canon_sha256":"8455f47f3b3d9eb693aaa048a1d58d4bca23a47090c0e4f76cd76b736325e538","abstract_canon_sha256":"8f6c5a4e907da9540ed1f1c0d05b1ee17aaa2c86cde169671a699ebe5d5535c9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T00:05:15.212793Z","signature_b64":"XuxKh0f/cRo5ZlVvBrTcgKM3Y3//x1pm1GBjngbIR5if5XeOrDlJ6Ase2twAP33BVTLmkK39OP+9/vPeyTJgBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d4bcd9c9b681cea0d287c9b99e0e7941c44117f5d574d3b87a011793317aa04e","last_reissued_at":"2026-05-28T00:05:15.211835Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T00:05:15.211835Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning in the Fisher Subspace: A Guided Initialization for LoRA Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hung-Yu Kao, Ying-Jia Lin, Zhi-Quan Feng","submitted_at":"2026-05-01T19:20:25Z","abstract_excerpt":"LoRA adapts large language models (LLMs) by restricting updates to low-rank subspaces of pre-trained weights. While this substantially reduces training cost, the effectiveness of adaptation critically depends on which subspace is chosen at initialization: a poor initialization that allocates capacity to task-irrelevant directions can severely hinder downstream performance. Existing initialization strategies primarily rely on the intrinsic properties of pre-trained weights, implicitly assuming that weight geometry alone reflects task relevance. However, such criteria overlook how the model inte"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Empirical results across diverse tasks and modalities demonstrate that data-aware initialization consistently and significantly improves downstream performance over existing approaches.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the curvature information induced by the downstream data distribution accurately identifies parameter directions whose perturbations most influence model predictions on the target objective.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Fisher information from target data provides a better criterion than weight geometry for choosing LoRA subspaces, yielding consistent performance gains on downstream tasks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4fe263de3d7f2221f758ef3207f7a5448cd5070aafd6700689665855ab8d5c6e"},"source":{"id":"2605.01046","kind":"arxiv","version":3},"verdict":{"id":"95a153b7-20d6-4c4b-97f8-c6e2655c7e5d","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-09T19:30:10.890636Z","strongest_claim":"Empirical results across diverse tasks and modalities demonstrate that data-aware initialization consistently and significantly improves downstream performance over existing approaches.","one_line_summary":"Fisher information from target data provides a better criterion than weight geometry for choosing LoRA subspaces, yielding consistent performance gains on downstream tasks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the curvature information induced by the downstream data distribution accurately identifies parameter directions whose perturbations most influence model predictions on the target objective.","pith_extraction_headline":"Using Fisher curvature from downstream data to initialize LoRA subspaces improves fine-tuning performance over weight-only methods."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.01046/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T18:39:26.354215Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T17:40:54.459980Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"a18f734de0fb534ad988031bd386a8993ac4714d14d0fd0bf72e76c0f01ca21b"},"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":"2605.01046","created_at":"2026-05-28T00:05:15.211960+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.01046v3","created_at":"2026-05-28T00:05:15.211960+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.01046","created_at":"2026-05-28T00:05:15.211960+00:00"},{"alias_kind":"pith_short_12","alias_value":"2S6NTSNWQHHK","created_at":"2026-05-28T00:05:15.211960+00:00"},{"alias_kind":"pith_short_16","alias_value":"2S6NTSNWQHHKBUUH","created_at":"2026-05-28T00:05:15.211960+00:00"},{"alias_kind":"pith_short_8","alias_value":"2S6NTSNW","created_at":"2026-05-28T00:05:15.211960+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/2S6NTSNWQHHKBUUHZG4Z4DTZIH","json":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH.json","graph_json":"https://pith.science/api/pith-number/2S6NTSNWQHHKBUUHZG4Z4DTZIH/graph.json","events_json":"https://pith.science/api/pith-number/2S6NTSNWQHHKBUUHZG4Z4DTZIH/events.json","paper":"https://pith.science/paper/2S6NTSNW"},"agent_actions":{"view_html":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH","download_json":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH.json","view_paper":"https://pith.science/paper/2S6NTSNW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.01046&json=true","fetch_graph":"https://pith.science/api/pith-number/2S6NTSNWQHHKBUUHZG4Z4DTZIH/graph.json","fetch_events":"https://pith.science/api/pith-number/2S6NTSNWQHHKBUUHZG4Z4DTZIH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/action/storage_attestation","attest_author":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/action/author_attestation","sign_citation":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/action/citation_signature","submit_replication":"https://pith.science/pith/2S6NTSNWQHHKBUUHZG4Z4DTZIH/action/replication_record"}},"created_at":"2026-05-28T00:05:15.211960+00:00","updated_at":"2026-05-28T00:05:15.211960+00:00"}