{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:J5AGE4GIQF2GNBIID6IXVE3LII","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"45cce6e35ed59ada4f5a31b882ff6434d169dc2587a8748eb47e178a18505745","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.OC","submitted_at":"2026-05-15T01:43:22Z","title_canon_sha256":"36c1edce1ac2cb48ee65483611a8698eddc85430de0bb7b6fffaf5ab624d7fbc"},"schema_version":"1.0","source":{"id":"2605.15522","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15522","created_at":"2026-05-20T00:01:03Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15522v1","created_at":"2026-05-20T00:01:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15522","created_at":"2026-05-20T00:01:03Z"},{"alias_kind":"pith_short_12","alias_value":"J5AGE4GIQF2G","created_at":"2026-05-20T00:01:03Z"},{"alias_kind":"pith_short_16","alias_value":"J5AGE4GIQF2GNBII","created_at":"2026-05-20T00:01:03Z"},{"alias_kind":"pith_short_8","alias_value":"J5AGE4GI","created_at":"2026-05-20T00:01:03Z"}],"graph_snapshots":[{"event_id":"sha256:df7e1349a90387cf5a52435e6ddff5f4aad63170bd2895d31555431565be7e1c","target":"graph","created_at":"2026-05-20T00:01:03Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"AdamW with clipped updates theoretically outperforms other popular stochastic optimization methods, such as SGD and AdaGrad, for convex stochastic generalized Lipschitz optimization problems."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"Objective functions belong to the newly introduced generalized Lipschitz class in which gradient norms are bounded by an affine function of the optimality gap."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Introduces generalized Lipschitz class and shows clipped AdamW outperforms SGD and AdaGrad for stochastic convex optimization under this and related assumptions."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"AdamW with clipped updates outperforms SGD and AdaGrad on convex stochastic problems with unbounded gradients."}],"snapshot_sha256":"1a11b3dd57969d76bd9f2221a527b49d8480ec234bb8b417cf1cc8a197cc783d"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T15:31:17.690711Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T15:20:34.474700Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"cited_work_retraction","ran_at":"2026-05-19T14:22:03.038115Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T14:21:54.045363Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"shingle_duplication","ran_at":"2026-05-19T13:49:41.841841Z","status":"skipped","version":"0.1.0"},{"findings_count":0,"name":"citation_quote_validity","ran_at":"2026-05-19T13:49:41.379071Z","status":"skipped","version":"0.1.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.627385Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.15522/integrity.json","findings":[],"snapshot_sha256":"a147eb8a3fb6396790e9f6f3f246dacf131b827a878ddc4fdab1fb44c81422f0","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Much of the existing theory on first-order non-smooth optimization is built on a restrictive assumption that the gradients of the objective function are uniformly bounded. We introduce a much more realistic class of generalized Lipschitz functions, where the gradient norms are bounded by an affine function of the optimality gap. We then ask a natural question: what algorithm achieves the best global convergence rates for solving convex stochastic generalized Lipschitz optimization problems? To address this, we develop a new convergence analysis for several existing algorithms and find that Ada","authors_text":"Dmitry Kovalev","cross_cats":["cs.LG"],"headline":"AdamW with clipped updates outperforms SGD and AdaGrad on convex stochastic problems with unbounded gradients.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.OC","submitted_at":"2026-05-15T01:43:22Z","title":"Stochastic Non-Smooth Convex Optimization with Unbounded Gradients"},"references":{"count":22,"internal_anchors":5,"resolved_work":22,"sample":[{"cited_arxiv_id":"1412.6980","doi":"","is_internal_anchor":true,"ref_index":1,"title":"Adam: A Method for Stochastic Optimization","work_id":"1910796d-9b52-4683-bf5c-de9632c1028b","year":null},{"cited_arxiv_id":"1711.05101","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Decoupled Weight Decay Regularization","work_id":"07ef7360-d385-4033-83f7-8384a6325204","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Why gradient clipping accelerates training: A theoretical justification for adaptivity","work_id":"fbf2f93c-f73e-413f-b24f-e6153c29bf03","year":1905},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Near-optimal methods for minimizing star-convex functions and beyond","work_id":"c69cc4c3-a26d-40c9-98be-6897c99bae26","year":1938},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Methods for convex (l_0, l_1)-smooth optimization: Clipping, acceleration, and adaptivity","work_id":"02737a27-7cd5-4b7a-a102-1c2c4d87a916","year":null}],"snapshot_sha256":"78d16cf0c3a5c950b97f185aadbb248c40a422e47582e75d3e07f33375deacc9"},"source":{"id":"2605.15522","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T15:11:36.275720Z","id":"1de4c210-3976-49b9-be99-5c1f364ce5b4","model_set":{"reader":"grok-4.3"},"one_line_summary":"Introduces generalized Lipschitz class and shows clipped AdamW outperforms SGD and AdaGrad for stochastic convex optimization under this and related assumptions.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"AdamW with clipped updates outperforms SGD and AdaGrad on convex stochastic problems with unbounded gradients.","strongest_claim":"AdamW with clipped updates theoretically outperforms other popular stochastic optimization methods, such as SGD and AdaGrad, for convex stochastic generalized Lipschitz optimization problems.","weakest_assumption":"Objective functions belong to the newly introduced generalized Lipschitz class in which gradient norms are bounded by an affine function of the optimality gap."}},"verdict_id":"1de4c210-3976-49b9-be99-5c1f364ce5b4"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:cd793ee1e03a1de6ff5c436aa1bcbc11bacf5ad63b70b07f0e29b255da3b5069","target":"record","created_at":"2026-05-20T00:01:03Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"45cce6e35ed59ada4f5a31b882ff6434d169dc2587a8748eb47e178a18505745","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.OC","submitted_at":"2026-05-15T01:43:22Z","title_canon_sha256":"36c1edce1ac2cb48ee65483611a8698eddc85430de0bb7b6fffaf5ab624d7fbc"},"schema_version":"1.0","source":{"id":"2605.15522","kind":"arxiv","version":1}},"canonical_sha256":"4f406270c881746685081f917a936b423c573ec24ec739af48937a6a2c20c7d4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4f406270c881746685081f917a936b423c573ec24ec739af48937a6a2c20c7d4","first_computed_at":"2026-05-20T00:01:03.043859Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:01:03.043859Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ykjhWXeU4rxrr+aC3KKmd9fc7Pp7RQzN1bQ2OZ/yJ4F0tU+b6RyPAFZvF0PfjwFFRmgk5VLsnfBYBtCRNW0OBQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:01:03.044951Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.15522","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:cd793ee1e03a1de6ff5c436aa1bcbc11bacf5ad63b70b07f0e29b255da3b5069","sha256:df7e1349a90387cf5a52435e6ddff5f4aad63170bd2895d31555431565be7e1c"],"state_sha256":"d59701edfb3729c2dad3339188306ca58466ba15e8667b7832f6be236bcde586"}