{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DN5NG37G7L3PL7UQUDGYOCYKN3","short_pith_number":"pith:DN5NG37G","schema_version":"1.0","canonical_sha256":"1b7ad36fe6faf6f5fe90a0cd870b0a6ed933904175274c37f5867971931dd615","source":{"kind":"arxiv","id":"2605.26547","version":1},"attestation_state":"computed","paper":{"title":"High-Probability Guarantees for Random Zeroth-Order Gradient Descent on Smooth Functions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"math.OC","authors_text":"Haishan Ye","submitted_at":"2026-05-26T04:56:24Z","abstract_excerpt":"Randomized zeroth-order methods are classically analyzed in expectation, but a black-box Markov conversion can give misleading high-probability guarantees, in particular by forcing the finite-difference smoothing radius to shrink with the confidence parameter. This paper gives a direct finite-horizon high-probability analysis of a two-query Gaussian finite-difference method for deterministic objectives with Lipschitz gradients. The method uses the classical two-point estimator together with the normalized stepsize \\(\\eta_t=1/(4L\\norm{\\bu_t}^2)\\). We prove that it finds an \\(\\varepsilon\\)-subop"},"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":"2605.26547","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"math.OC","submitted_at":"2026-05-26T04:56:24Z","cross_cats_sorted":[],"title_canon_sha256":"877c4cd60642c7bd3d241f11f514d0e6078d47f3d41d7193afe7363174c4ad19","abstract_canon_sha256":"3ff2c6602a29fa23c5c0ba006edbb88a3895e4290a6bd797c181127547f75aeb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:05:25.854945Z","signature_b64":"s7QBTXJIbM8mGUGOTj4e87HJkWNZMbu+BLSWxkjym9Z8nC62Uc7nOxUfijTDb02pkJxx7BGxgE2Ami+wcXzyCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1b7ad36fe6faf6f5fe90a0cd870b0a6ed933904175274c37f5867971931dd615","last_reissued_at":"2026-05-27T01:05:25.854497Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:05:25.854497Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"High-Probability Guarantees for Random Zeroth-Order Gradient Descent on Smooth Functions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"math.OC","authors_text":"Haishan Ye","submitted_at":"2026-05-26T04:56:24Z","abstract_excerpt":"Randomized zeroth-order methods are classically analyzed in expectation, but a black-box Markov conversion can give misleading high-probability guarantees, in particular by forcing the finite-difference smoothing radius to shrink with the confidence parameter. This paper gives a direct finite-horizon high-probability analysis of a two-query Gaussian finite-difference method for deterministic objectives with Lipschitz gradients. The method uses the classical two-point estimator together with the normalized stepsize \\(\\eta_t=1/(4L\\norm{\\bu_t}^2)\\). We prove that it finds an \\(\\varepsilon\\)-subop"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.26547","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/2605.26547/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":"2605.26547","created_at":"2026-05-27T01:05:25.854563+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.26547v1","created_at":"2026-05-27T01:05:25.854563+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.26547","created_at":"2026-05-27T01:05:25.854563+00:00"},{"alias_kind":"pith_short_12","alias_value":"DN5NG37G7L3P","created_at":"2026-05-27T01:05:25.854563+00:00"},{"alias_kind":"pith_short_16","alias_value":"DN5NG37G7L3PL7UQ","created_at":"2026-05-27T01:05:25.854563+00:00"},{"alias_kind":"pith_short_8","alias_value":"DN5NG37G","created_at":"2026-05-27T01:05:25.854563+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2606.20446","citing_title":"High-Probability Last-Iterate Guarantees for Two-Point Gaussian Zeroth-Order Stochastic Gradient Descent","ref_index":20,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3","json":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3.json","graph_json":"https://pith.science/api/pith-number/DN5NG37G7L3PL7UQUDGYOCYKN3/graph.json","events_json":"https://pith.science/api/pith-number/DN5NG37G7L3PL7UQUDGYOCYKN3/events.json","paper":"https://pith.science/paper/DN5NG37G"},"agent_actions":{"view_html":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3","download_json":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3.json","view_paper":"https://pith.science/paper/DN5NG37G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.26547&json=true","fetch_graph":"https://pith.science/api/pith-number/DN5NG37G7L3PL7UQUDGYOCYKN3/graph.json","fetch_events":"https://pith.science/api/pith-number/DN5NG37G7L3PL7UQUDGYOCYKN3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3/action/storage_attestation","attest_author":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3/action/author_attestation","sign_citation":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3/action/citation_signature","submit_replication":"https://pith.science/pith/DN5NG37G7L3PL7UQUDGYOCYKN3/action/replication_record"}},"created_at":"2026-05-27T01:05:25.854563+00:00","updated_at":"2026-05-27T01:05:25.854563+00:00"}