{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:HYHMAPFKKWUSF4CGEXJXBX274V","short_pith_number":"pith:HYHMAPFK","schema_version":"1.0","canonical_sha256":"3e0ec03caa55a922f04625d370df5fe547306b7afc48ee4d1582fd4d505c7f20","source":{"kind":"arxiv","id":"2502.06820","version":2},"attestation_state":"computed","paper":{"title":"LoCA: Location-Aware Cosine Adaptation for Parameter-Efficient Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Changliang Zou, Jingjing Li, Ke Lu, Liuhua Peng, Mingming Gong, Tingjin Chu, Yinjie Min, Zhekai Du","submitted_at":"2025-02-05T04:14:34Z","abstract_excerpt":"Low-rank adaptation (LoRA) has become a prevalent method for adapting pre-trained large language models to downstream tasks. However, the simple low-rank decomposition form may constrain the hypothesis space. To address this limitation, we introduce Location-aware Cosine Adaptation (LoCA), a novel frequency-domain parameter-efficient fine-tuning method based on inverse Discrete Cosine Transform (iDCT) with selective locations of learnable components. We begin with a comprehensive theoretical comparison between frequency-domain and low-rank decompositions for fine-tuning pre-trained large model"},"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":"2502.06820","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-05T04:14:34Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"534ce725af3608a261708d0f313933e7d3b9a9ed289d2853fbc3c2b9e829a7ba","abstract_canon_sha256":"9519833c4a8546da646dee6c8ba6c8205b29f481bb4076bd12ba327869dff5b1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:55:30.312678Z","signature_b64":"yKjHzKCtTeRPFJan6ZnFTgexZb1UF7ynWa6DaZpe7d0F6eGpl5zPwP3NHK1wjwm/+PXw7hRdpbcdJBcFPPEjDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3e0ec03caa55a922f04625d370df5fe547306b7afc48ee4d1582fd4d505c7f20","last_reissued_at":"2026-07-05T10:55:30.312106Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:55:30.312106Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LoCA: Location-Aware Cosine Adaptation for Parameter-Efficient Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Changliang Zou, Jingjing Li, Ke Lu, Liuhua Peng, Mingming Gong, Tingjin Chu, Yinjie Min, Zhekai Du","submitted_at":"2025-02-05T04:14:34Z","abstract_excerpt":"Low-rank adaptation (LoRA) has become a prevalent method for adapting pre-trained large language models to downstream tasks. However, the simple low-rank decomposition form may constrain the hypothesis space. To address this limitation, we introduce Location-aware Cosine Adaptation (LoCA), a novel frequency-domain parameter-efficient fine-tuning method based on inverse Discrete Cosine Transform (iDCT) with selective locations of learnable components. We begin with a comprehensive theoretical comparison between frequency-domain and low-rank decompositions for fine-tuning pre-trained large model"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.06820","kind":"arxiv","version":2},"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/2502.06820/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":"2502.06820","created_at":"2026-07-05T10:55:30.312168+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.06820v2","created_at":"2026-07-05T10:55:30.312168+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.06820","created_at":"2026-07-05T10:55:30.312168+00:00"},{"alias_kind":"pith_short_12","alias_value":"HYHMAPFKKWUS","created_at":"2026-07-05T10:55:30.312168+00:00"},{"alias_kind":"pith_short_16","alias_value":"HYHMAPFKKWUSF4CG","created_at":"2026-07-05T10:55:30.312168+00:00"},{"alias_kind":"pith_short_8","alias_value":"HYHMAPFK","created_at":"2026-07-05T10:55:30.312168+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.02842","citing_title":"Spectral-Progressive Thought Flow for Lightweight Multimodal Reasoning","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V","json":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V.json","graph_json":"https://pith.science/api/pith-number/HYHMAPFKKWUSF4CGEXJXBX274V/graph.json","events_json":"https://pith.science/api/pith-number/HYHMAPFKKWUSF4CGEXJXBX274V/events.json","paper":"https://pith.science/paper/HYHMAPFK"},"agent_actions":{"view_html":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V","download_json":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V.json","view_paper":"https://pith.science/paper/HYHMAPFK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.06820&json=true","fetch_graph":"https://pith.science/api/pith-number/HYHMAPFKKWUSF4CGEXJXBX274V/graph.json","fetch_events":"https://pith.science/api/pith-number/HYHMAPFKKWUSF4CGEXJXBX274V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V/action/storage_attestation","attest_author":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V/action/author_attestation","sign_citation":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V/action/citation_signature","submit_replication":"https://pith.science/pith/HYHMAPFKKWUSF4CGEXJXBX274V/action/replication_record"}},"created_at":"2026-07-05T10:55:30.312168+00:00","updated_at":"2026-07-05T10:55:30.312168+00:00"}