{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NLZV37EIOKBDW7LQPMIWO2FC5M","short_pith_number":"pith:NLZV37EI","schema_version":"1.0","canonical_sha256":"6af35dfc8872823b7d707b116768a2eb1d3f264f1daa3dc45ec4c4625815cd75","source":{"kind":"arxiv","id":"2505.12579","version":1},"attestation_state":"computed","paper":{"title":"Adaptive parameter-efficient fine-tuning via Hessian-informed subset selection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Shiyun Xu, Zhiqi Bu","submitted_at":"2025-05-18T23:45:50Z","abstract_excerpt":"Parameter-efficient fine-tuning (PEFT) is a highly effective approach for adapting large pre-trained models to downstream tasks with minimal computational overhead. At the core, PEFT methods freeze most parameters and only trains a small subset (say $<0.1\\%$ of total parameters). Notably, different PEFT methods select different subsets, resulting in varying levels of performance. This variation prompts a key question: how to effectively select the most influential subset to train?\n  We formulate the subset selection as a multi-task problem: maximizing the performance and minimizing the number "},"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":"2505.12579","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-18T23:45:50Z","cross_cats_sorted":[],"title_canon_sha256":"4edbbb4b8e25959ac4ff6ddb32023a5523fc9b547ac3b97e8723082d99f0c1aa","abstract_canon_sha256":"4392e67b7e5bcc52d9b2ed6ef11d5e2ff84e1fb4b36b54c08e7e28abf8137c55"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:05:06.354289Z","signature_b64":"jr8QVZgZC+nqCBNe4ntZAUj9efBwWOutwqBBOz3otCtoHBaHChA4NY2RZKcd8CtKoAbpjuxi+QEz6zuO3lDJAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6af35dfc8872823b7d707b116768a2eb1d3f264f1daa3dc45ec4c4625815cd75","last_reissued_at":"2026-07-05T11:05:06.353717Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:05:06.353717Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adaptive parameter-efficient fine-tuning via Hessian-informed subset selection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Shiyun Xu, Zhiqi Bu","submitted_at":"2025-05-18T23:45:50Z","abstract_excerpt":"Parameter-efficient fine-tuning (PEFT) is a highly effective approach for adapting large pre-trained models to downstream tasks with minimal computational overhead. At the core, PEFT methods freeze most parameters and only trains a small subset (say $<0.1\\%$ of total parameters). Notably, different PEFT methods select different subsets, resulting in varying levels of performance. This variation prompts a key question: how to effectively select the most influential subset to train?\n  We formulate the subset selection as a multi-task problem: maximizing the performance and minimizing the number "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.12579","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/2505.12579/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":"2505.12579","created_at":"2026-07-05T11:05:06.353778+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.12579v1","created_at":"2026-07-05T11:05:06.353778+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.12579","created_at":"2026-07-05T11:05:06.353778+00:00"},{"alias_kind":"pith_short_12","alias_value":"NLZV37EIOKBD","created_at":"2026-07-05T11:05:06.353778+00:00"},{"alias_kind":"pith_short_16","alias_value":"NLZV37EIOKBDW7LQ","created_at":"2026-07-05T11:05:06.353778+00:00"},{"alias_kind":"pith_short_8","alias_value":"NLZV37EI","created_at":"2026-07-05T11:05:06.353778+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.16191","citing_title":"GEM: A Scale-Aware and Distribution-Sensitive Sparse Fine-Tuning Framework for Effective Downstream Adaptation","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M","json":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M.json","graph_json":"https://pith.science/api/pith-number/NLZV37EIOKBDW7LQPMIWO2FC5M/graph.json","events_json":"https://pith.science/api/pith-number/NLZV37EIOKBDW7LQPMIWO2FC5M/events.json","paper":"https://pith.science/paper/NLZV37EI"},"agent_actions":{"view_html":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M","download_json":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M.json","view_paper":"https://pith.science/paper/NLZV37EI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.12579&json=true","fetch_graph":"https://pith.science/api/pith-number/NLZV37EIOKBDW7LQPMIWO2FC5M/graph.json","fetch_events":"https://pith.science/api/pith-number/NLZV37EIOKBDW7LQPMIWO2FC5M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M/action/storage_attestation","attest_author":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M/action/author_attestation","sign_citation":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M/action/citation_signature","submit_replication":"https://pith.science/pith/NLZV37EIOKBDW7LQPMIWO2FC5M/action/replication_record"}},"created_at":"2026-07-05T11:05:06.353778+00:00","updated_at":"2026-07-05T11:05:06.353778+00:00"}