{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:4O3PH523DYRYVI42HSFTWLJWNM","short_pith_number":"pith:4O3PH523","schema_version":"1.0","canonical_sha256":"e3b6f3f75b1e238aa39a3c8b3b2d366b224d826fee6bfbc50ba007bcefb2571c","source":{"kind":"arxiv","id":"1803.07300","version":3},"attestation_state":"computed","paper":{"title":"Risk and parameter convergence of logistic regression","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Matus Telgarsky, Ziwei Ji","submitted_at":"2018-03-20T08:47:27Z","abstract_excerpt":"Gradient descent, when applied to the task of logistic regression, outputs iterates which are biased to follow a unique ray defined by the data. The direction of this ray is the maximum margin predictor of a maximal linearly separable subset of the data; the gradient descent iterates converge to this ray in direction at the rate $\\mathcal{O}(\\ln\\ln t / \\ln t)$. The ray does not pass through the origin in general, and its offset is the bounded global optimum of the risk over the remaining data; gradient descent recovers this offset at a rate $\\mathcal{O}((\\ln t)^2 / \\sqrt{t})$."},"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":"1803.07300","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-03-20T08:47:27Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"4079d05ce1b99ee278d429575aa352c28ed12039f837124c618b15213e47640e","abstract_canon_sha256":"11efd6ec67cf80fc6e669d81af35b469d550ed9cf2912b050bba4fcbc77acf60"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:43:51.345355Z","signature_b64":"rBsGAzhwGMr8JO9wEq/NuCjPr/9ytqEK39usQzv0xIqrwoCIlef979KD8m9vuLf4a4m03JTDvdMd1bv9vX2cCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e3b6f3f75b1e238aa39a3c8b3b2d366b224d826fee6bfbc50ba007bcefb2571c","last_reissued_at":"2026-05-17T23:43:51.344803Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:43:51.344803Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Risk and parameter convergence of logistic regression","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Matus Telgarsky, Ziwei Ji","submitted_at":"2018-03-20T08:47:27Z","abstract_excerpt":"Gradient descent, when applied to the task of logistic regression, outputs iterates which are biased to follow a unique ray defined by the data. The direction of this ray is the maximum margin predictor of a maximal linearly separable subset of the data; the gradient descent iterates converge to this ray in direction at the rate $\\mathcal{O}(\\ln\\ln t / \\ln t)$. The ray does not pass through the origin in general, and its offset is the bounded global optimum of the risk over the remaining data; gradient descent recovers this offset at a rate $\\mathcal{O}((\\ln t)^2 / \\sqrt{t})$."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1803.07300","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"1803.07300","created_at":"2026-05-17T23:43:51.344899+00:00"},{"alias_kind":"arxiv_version","alias_value":"1803.07300v3","created_at":"2026-05-17T23:43:51.344899+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1803.07300","created_at":"2026-05-17T23:43:51.344899+00:00"},{"alias_kind":"pith_short_12","alias_value":"4O3PH523DYRY","created_at":"2026-05-18T12:32:05.422762+00:00"},{"alias_kind":"pith_short_16","alias_value":"4O3PH523DYRYVI42","created_at":"2026-05-18T12:32:05.422762+00:00"},{"alias_kind":"pith_short_8","alias_value":"4O3PH523","created_at":"2026-05-18T12:32:05.422762+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2512.11587","citing_title":"Gradient Descent as a Perceptron Algorithm: Understanding Dynamics and Implicit Acceleration","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2602.01642","citing_title":"The Effect of Mini-Batch Noise on the Implicit Bias of Adam","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2604.02920","citing_title":"Efficient Logistic Regression with Mixture of Sigmoids","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM","json":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM.json","graph_json":"https://pith.science/api/pith-number/4O3PH523DYRYVI42HSFTWLJWNM/graph.json","events_json":"https://pith.science/api/pith-number/4O3PH523DYRYVI42HSFTWLJWNM/events.json","paper":"https://pith.science/paper/4O3PH523"},"agent_actions":{"view_html":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM","download_json":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM.json","view_paper":"https://pith.science/paper/4O3PH523","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1803.07300&json=true","fetch_graph":"https://pith.science/api/pith-number/4O3PH523DYRYVI42HSFTWLJWNM/graph.json","fetch_events":"https://pith.science/api/pith-number/4O3PH523DYRYVI42HSFTWLJWNM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM/action/storage_attestation","attest_author":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM/action/author_attestation","sign_citation":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM/action/citation_signature","submit_replication":"https://pith.science/pith/4O3PH523DYRYVI42HSFTWLJWNM/action/replication_record"}},"created_at":"2026-05-17T23:43:51.344899+00:00","updated_at":"2026-05-17T23:43:51.344899+00:00"}