{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:PAKMTQSSLKX64KASTUM45UZJJS","short_pith_number":"pith:PAKMTQSS","schema_version":"1.0","canonical_sha256":"7814c9c2525aafee28129d19ced3294c81db407b953b99e1b0f5dcd8b278c41d","source":{"kind":"arxiv","id":"2206.06900","version":1},"attestation_state":"computed","paper":{"title":"Grad-GradaGrad? A Non-Monotone Adaptive Stochastic Gradient Method","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaron Defazio, Baoyu Zhou, Lin Xiao","submitted_at":"2022-06-14T14:55:27Z","abstract_excerpt":"The classical AdaGrad method adapts the learning rate by dividing by the square root of a sum of squared gradients. Because this sum on the denominator is increasing, the method can only decrease step sizes over time, and requires a learning rate scaling hyper-parameter to be carefully tuned. To overcome this restriction, we introduce GradaGrad, a method in the same family that naturally grows or shrinks the learning rate based on a different accumulation in the denominator, one that can both increase and decrease. We show that it obeys a similar convergence rate as AdaGrad and demonstrate its"},"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":"2206.06900","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-06-14T14:55:27Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"50fea321e3ff5161ee26da7ced04a62b3bf62b7a876e848c5e43e1f31b7771bf","abstract_canon_sha256":"fa4cd8e2b933fb57dc06a0040a4af51fba6d6a32e246f7b12cb376d2dac09a98"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:31:47.037237Z","signature_b64":"4H63ZNd1Q4sS+XsYc4X9TgiAoYpLQB6Ka9Pxi9bP21PA4yL/KRoinusKakQvSDJ/bjmX3nZfds6hD+CJo1uhBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7814c9c2525aafee28129d19ced3294c81db407b953b99e1b0f5dcd8b278c41d","last_reissued_at":"2026-07-05T04:31:47.036732Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:31:47.036732Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Grad-GradaGrad? A Non-Monotone Adaptive Stochastic Gradient Method","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaron Defazio, Baoyu Zhou, Lin Xiao","submitted_at":"2022-06-14T14:55:27Z","abstract_excerpt":"The classical AdaGrad method adapts the learning rate by dividing by the square root of a sum of squared gradients. Because this sum on the denominator is increasing, the method can only decrease step sizes over time, and requires a learning rate scaling hyper-parameter to be carefully tuned. To overcome this restriction, we introduce GradaGrad, a method in the same family that naturally grows or shrinks the learning rate based on a different accumulation in the denominator, one that can both increase and decrease. We show that it obeys a similar convergence rate as AdaGrad and demonstrate its"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2206.06900","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/2206.06900/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":"2206.06900","created_at":"2026-07-05T04:31:47.036795+00:00"},{"alias_kind":"arxiv_version","alias_value":"2206.06900v1","created_at":"2026-07-05T04:31:47.036795+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2206.06900","created_at":"2026-07-05T04:31:47.036795+00:00"},{"alias_kind":"pith_short_12","alias_value":"PAKMTQSSLKX6","created_at":"2026-07-05T04:31:47.036795+00:00"},{"alias_kind":"pith_short_16","alias_value":"PAKMTQSSLKX64KAS","created_at":"2026-07-05T04:31:47.036795+00:00"},{"alias_kind":"pith_short_8","alias_value":"PAKMTQSS","created_at":"2026-07-05T04:31:47.036795+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.15522","citing_title":"Stochastic Non-Smooth Convex Optimization with Unbounded Gradients","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS","json":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS.json","graph_json":"https://pith.science/api/pith-number/PAKMTQSSLKX64KASTUM45UZJJS/graph.json","events_json":"https://pith.science/api/pith-number/PAKMTQSSLKX64KASTUM45UZJJS/events.json","paper":"https://pith.science/paper/PAKMTQSS"},"agent_actions":{"view_html":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS","download_json":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS.json","view_paper":"https://pith.science/paper/PAKMTQSS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2206.06900&json=true","fetch_graph":"https://pith.science/api/pith-number/PAKMTQSSLKX64KASTUM45UZJJS/graph.json","fetch_events":"https://pith.science/api/pith-number/PAKMTQSSLKX64KASTUM45UZJJS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS/action/storage_attestation","attest_author":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS/action/author_attestation","sign_citation":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS/action/citation_signature","submit_replication":"https://pith.science/pith/PAKMTQSSLKX64KASTUM45UZJJS/action/replication_record"}},"created_at":"2026-07-05T04:31:47.036795+00:00","updated_at":"2026-07-05T04:31:47.036795+00:00"}