{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:DCZIR6LLIQM2JBY7JAVUFQDKCH","short_pith_number":"pith:DCZIR6LL","schema_version":"1.0","canonical_sha256":"18b288f96b4419a4871f482b42c06a11d5aa779be1e86f82d027ffa2dc888fe0","source":{"kind":"arxiv","id":"2412.00071","version":2},"attestation_state":"computed","paper":{"title":"COAP: Memory-Efficient Training with Correlation-Aware Gradient Projection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Bo Yuan, Jing Liu, Jinqi Xiao, Linjie Luo, Qing Yan, Shen Sang, Tiancheng Zhi, Yuqian Zhang","submitted_at":"2024-11-26T03:50:52Z","abstract_excerpt":"Training large-scale neural networks in vision, and multimodal domains demands substantial memory resources, primarily due to the storage of optimizer states. While LoRA, a popular parameter-efficient method, reduces memory usage, it often suffers from suboptimal performance due to the constraints of low-rank updates. Low-rank gradient projection methods (e.g., GaLore, Flora) reduce optimizer memory by projecting gradients and moment estimates into low-rank spaces via singular value decomposition or random projection. However, they fail to account for inter-projection correlation, causing perf"},"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":"2412.00071","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-11-26T03:50:52Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CV"],"title_canon_sha256":"eb5f7164b6b0a2cf6829bc68f180756f790dcd3357fe917254c6755c026158cd","abstract_canon_sha256":"a8bf59385a260ed8801683322b9455bdd69cf81dca9d9921c31cec9e0a7c25e0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:29:27.507635Z","signature_b64":"ed7TIUDE1cWEDAtLdKkof/Vfs3J6zDXafzUlsuz3s7WFbhbs9WQmeXdJaIAWSbNfZ1JBk1+VdShPgU8GPOL6AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"18b288f96b4419a4871f482b42c06a11d5aa779be1e86f82d027ffa2dc888fe0","last_reissued_at":"2026-07-05T10:29:27.506683Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:29:27.506683Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"COAP: Memory-Efficient Training with Correlation-Aware Gradient Projection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Bo Yuan, Jing Liu, Jinqi Xiao, Linjie Luo, Qing Yan, Shen Sang, Tiancheng Zhi, Yuqian Zhang","submitted_at":"2024-11-26T03:50:52Z","abstract_excerpt":"Training large-scale neural networks in vision, and multimodal domains demands substantial memory resources, primarily due to the storage of optimizer states. While LoRA, a popular parameter-efficient method, reduces memory usage, it often suffers from suboptimal performance due to the constraints of low-rank updates. Low-rank gradient projection methods (e.g., GaLore, Flora) reduce optimizer memory by projecting gradients and moment estimates into low-rank spaces via singular value decomposition or random projection. However, they fail to account for inter-projection correlation, causing perf"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.00071","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/2412.00071/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":"2412.00071","created_at":"2026-07-05T10:29:27.506793+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.00071v2","created_at":"2026-07-05T10:29:27.506793+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.00071","created_at":"2026-07-05T10:29:27.506793+00:00"},{"alias_kind":"pith_short_12","alias_value":"DCZIR6LLIQM2","created_at":"2026-07-05T10:29:27.506793+00:00"},{"alias_kind":"pith_short_16","alias_value":"DCZIR6LLIQM2JBY7","created_at":"2026-07-05T10:29:27.506793+00:00"},{"alias_kind":"pith_short_8","alias_value":"DCZIR6LL","created_at":"2026-07-05T10:29:27.506793+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2512.21788","citing_title":"InstructMoLE: Instruction-Guided Mixture of Low-rank Experts for Multi-Conditional Image Generation","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH","json":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH.json","graph_json":"https://pith.science/api/pith-number/DCZIR6LLIQM2JBY7JAVUFQDKCH/graph.json","events_json":"https://pith.science/api/pith-number/DCZIR6LLIQM2JBY7JAVUFQDKCH/events.json","paper":"https://pith.science/paper/DCZIR6LL"},"agent_actions":{"view_html":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH","download_json":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH.json","view_paper":"https://pith.science/paper/DCZIR6LL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.00071&json=true","fetch_graph":"https://pith.science/api/pith-number/DCZIR6LLIQM2JBY7JAVUFQDKCH/graph.json","fetch_events":"https://pith.science/api/pith-number/DCZIR6LLIQM2JBY7JAVUFQDKCH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH/action/storage_attestation","attest_author":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH/action/author_attestation","sign_citation":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH/action/citation_signature","submit_replication":"https://pith.science/pith/DCZIR6LLIQM2JBY7JAVUFQDKCH/action/replication_record"}},"created_at":"2026-07-05T10:29:27.506793+00:00","updated_at":"2026-07-05T10:29:27.506793+00:00"}