{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3G5R3WBWPB7RWMF5JACFIMQTIO","short_pith_number":"pith:3G5R3WBW","schema_version":"1.0","canonical_sha256":"d9bb1dd836787f1b30bd48045432134397604ffaf650cd7b0c1946edfe6a6d12","source":{"kind":"arxiv","id":"2502.01014","version":2},"attestation_state":"computed","paper":{"title":"Refining Adaptive Zeroth-Order Optimization at Ease","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Kun He, Qixin Zhang, Yao Shu, Zhongxiang Dai","submitted_at":"2025-02-03T03:10:44Z","abstract_excerpt":"Recently, zeroth-order (ZO) optimization plays an essential role in scenarios where gradient information is inaccessible or unaffordable, such as black-box systems and resource-constrained environments. While existing adaptive methods such as ZO-AdaMM have shown promise, they are fundamentally limited by their underutilization of moment information during optimization, usually resulting in underperforming convergence. To overcome these limitations, this paper introduces Refined Adaptive Zeroth-Order Optimization (R-AdaZO). Specifically, we first show the untapped variance reduction effect of f"},"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":"2502.01014","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-03T03:10:44Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ef318e41d0a491a0735494fded5380dafb8c771c3659c007414d7f605ac11fc2","abstract_canon_sha256":"906aab6c4b13199915258e8efa6361acc06e4289a13a2c0781879e01f7b3ad30"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:17:40.395611Z","signature_b64":"AxWFtTlzEw0xtufE6K2A4bzVqx9sAdMqgM9EOALJ+rKLRA1WrNr9/hm637rexPGhu5eT/HUBvq0OA+IunibeCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d9bb1dd836787f1b30bd48045432134397604ffaf650cd7b0c1946edfe6a6d12","last_reissued_at":"2026-07-05T11:17:40.395124Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:17:40.395124Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Refining Adaptive Zeroth-Order Optimization at Ease","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Kun He, Qixin Zhang, Yao Shu, Zhongxiang Dai","submitted_at":"2025-02-03T03:10:44Z","abstract_excerpt":"Recently, zeroth-order (ZO) optimization plays an essential role in scenarios where gradient information is inaccessible or unaffordable, such as black-box systems and resource-constrained environments. While existing adaptive methods such as ZO-AdaMM have shown promise, they are fundamentally limited by their underutilization of moment information during optimization, usually resulting in underperforming convergence. To overcome these limitations, this paper introduces Refined Adaptive Zeroth-Order Optimization (R-AdaZO). Specifically, we first show the untapped variance reduction effect of f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.01014","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/2502.01014/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":"2502.01014","created_at":"2026-07-05T11:17:40.395179+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.01014v2","created_at":"2026-07-05T11:17:40.395179+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.01014","created_at":"2026-07-05T11:17:40.395179+00:00"},{"alias_kind":"pith_short_12","alias_value":"3G5R3WBWPB7R","created_at":"2026-07-05T11:17:40.395179+00:00"},{"alias_kind":"pith_short_16","alias_value":"3G5R3WBWPB7RWMF5","created_at":"2026-07-05T11:17:40.395179+00:00"},{"alias_kind":"pith_short_8","alias_value":"3G5R3WBW","created_at":"2026-07-05T11:17:40.395179+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.09034","citing_title":"Accelerating Zeroth-Order Spectral Optimization with Partial Orthogonalization from Power Iteration","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2510.00761","citing_title":"Downgrade to Upgrade: Optimizer Simplification Enhances Robustness in LLM Unlearning","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09034","citing_title":"Accelerating Zeroth-Order Spectral Optimization with Partial Orthogonalization from Power Iteration","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO","json":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO.json","graph_json":"https://pith.science/api/pith-number/3G5R3WBWPB7RWMF5JACFIMQTIO/graph.json","events_json":"https://pith.science/api/pith-number/3G5R3WBWPB7RWMF5JACFIMQTIO/events.json","paper":"https://pith.science/paper/3G5R3WBW"},"agent_actions":{"view_html":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO","download_json":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO.json","view_paper":"https://pith.science/paper/3G5R3WBW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.01014&json=true","fetch_graph":"https://pith.science/api/pith-number/3G5R3WBWPB7RWMF5JACFIMQTIO/graph.json","fetch_events":"https://pith.science/api/pith-number/3G5R3WBWPB7RWMF5JACFIMQTIO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO/action/storage_attestation","attest_author":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO/action/author_attestation","sign_citation":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO/action/citation_signature","submit_replication":"https://pith.science/pith/3G5R3WBWPB7RWMF5JACFIMQTIO/action/replication_record"}},"created_at":"2026-07-05T11:17:40.395179+00:00","updated_at":"2026-07-05T11:17:40.395179+00:00"}