{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:QVHHWI6SCO2GA4A5L4C5DPEYFM","short_pith_number":"pith:QVHHWI6S","schema_version":"1.0","canonical_sha256":"854e7b23d213b460701d5f05d1bc982b1d386865b4ab3678c4371dc628973c09","source":{"kind":"arxiv","id":"2103.04902","version":2},"attestation_state":"computed","paper":{"title":"Stochasticity helps to navigate rough landscapes: comparing gradient-descent-based algorithms in the phase retrieval problem","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.ST","stat.ML","stat.TH"],"primary_cat":"cond-mat.dis-nn","authors_text":"Francesca Mignacco, Lenka Zdeborov\\'a, Pierfrancesco Urbani","submitted_at":"2021-03-08T17:06:18Z","abstract_excerpt":"In this paper we investigate how gradient-based algorithms such as gradient descent, (multi-pass) stochastic gradient descent, its persistent variant, and the Langevin algorithm navigate non-convex loss-landscapes and which of them is able to reach the best generalization error at limited sample complexity. We consider the loss landscape of the high-dimensional phase retrieval problem as a prototypical highly non-convex example. We observe that for phase retrieval the stochastic variants of gradient descent are able to reach perfect generalization for regions of control parameters where the gr"},"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":"2103.04902","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cond-mat.dis-nn","submitted_at":"2021-03-08T17:06:18Z","cross_cats_sorted":["cs.LG","math.ST","stat.ML","stat.TH"],"title_canon_sha256":"997662b1cb9b3b4d8e0346682533b099033ee7bb1ae4b3c7603c44c1fcc0719e","abstract_canon_sha256":"c3413ae968d014829a598e70a88acb0647f17abd6611186843c3113fd3faddf6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:06:41.416360Z","signature_b64":"hBLTypamYVKJtMFOB+IoRQU1nNSg+oqsgZCfguQpxo28JYeeumP7d4YANQmgpDQ0ttHiAyN2XeVUA/G2loctCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"854e7b23d213b460701d5f05d1bc982b1d386865b4ab3678c4371dc628973c09","last_reissued_at":"2026-07-05T04:06:41.415930Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:06:41.415930Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stochasticity helps to navigate rough landscapes: comparing gradient-descent-based algorithms in the phase retrieval problem","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.ST","stat.ML","stat.TH"],"primary_cat":"cond-mat.dis-nn","authors_text":"Francesca Mignacco, Lenka Zdeborov\\'a, Pierfrancesco Urbani","submitted_at":"2021-03-08T17:06:18Z","abstract_excerpt":"In this paper we investigate how gradient-based algorithms such as gradient descent, (multi-pass) stochastic gradient descent, its persistent variant, and the Langevin algorithm navigate non-convex loss-landscapes and which of them is able to reach the best generalization error at limited sample complexity. We consider the loss landscape of the high-dimensional phase retrieval problem as a prototypical highly non-convex example. We observe that for phase retrieval the stochastic variants of gradient descent are able to reach perfect generalization for regions of control parameters where the gr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.04902","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/2103.04902/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":"2103.04902","created_at":"2026-07-05T04:06:41.415987+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.04902v2","created_at":"2026-07-05T04:06:41.415987+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.04902","created_at":"2026-07-05T04:06:41.415987+00:00"},{"alias_kind":"pith_short_12","alias_value":"QVHHWI6SCO2G","created_at":"2026-07-05T04:06:41.415987+00:00"},{"alias_kind":"pith_short_16","alias_value":"QVHHWI6SCO2GA4A5","created_at":"2026-07-05T04:06:41.415987+00:00"},{"alias_kind":"pith_short_8","alias_value":"QVHHWI6S","created_at":"2026-07-05T04:06:41.415987+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM","json":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM.json","graph_json":"https://pith.science/api/pith-number/QVHHWI6SCO2GA4A5L4C5DPEYFM/graph.json","events_json":"https://pith.science/api/pith-number/QVHHWI6SCO2GA4A5L4C5DPEYFM/events.json","paper":"https://pith.science/paper/QVHHWI6S"},"agent_actions":{"view_html":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM","download_json":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM.json","view_paper":"https://pith.science/paper/QVHHWI6S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.04902&json=true","fetch_graph":"https://pith.science/api/pith-number/QVHHWI6SCO2GA4A5L4C5DPEYFM/graph.json","fetch_events":"https://pith.science/api/pith-number/QVHHWI6SCO2GA4A5L4C5DPEYFM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM/action/storage_attestation","attest_author":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM/action/author_attestation","sign_citation":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM/action/citation_signature","submit_replication":"https://pith.science/pith/QVHHWI6SCO2GA4A5L4C5DPEYFM/action/replication_record"}},"created_at":"2026-07-05T04:06:41.415987+00:00","updated_at":"2026-07-05T04:06:41.415987+00:00"}