{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:LQELKQKXEFWCG7TJ55B4V2XLCJ","short_pith_number":"pith:LQELKQKX","schema_version":"1.0","canonical_sha256":"5c08b54157216c237e69ef43caeaeb127dce9bcce02d8afd6f4200be312b346c","source":{"kind":"arxiv","id":"1801.03744","version":3},"attestation_state":"computed","paper":{"title":"Which Neural Net Architectures Give Rise To Exploding and Vanishing Gradients?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.PR","math.ST","stat.TH"],"primary_cat":"stat.ML","authors_text":"Boris Hanin","submitted_at":"2018-01-11T13:17:22Z","abstract_excerpt":"We give a rigorous analysis of the statistical behavior of gradients in a randomly initialized fully connected network N with ReLU activations. Our results show that the empirical variance of the squares of the entries in the input-output Jacobian of N is exponential in a simple architecture-dependent constant beta, given by the sum of the reciprocals of the hidden layer widths. When beta is large, the gradients computed by N at initialization vary wildly. Our approach complements the mean field theory analysis of random networks. From this point of view, we rigorously compute finite width cor"},"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":"1801.03744","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-01-11T13:17:22Z","cross_cats_sorted":["cs.LG","math.PR","math.ST","stat.TH"],"title_canon_sha256":"9c1b99ab75f9a8356b87a57ea69bc8fe8b6908dc0f37639898f3bb2a934d9c6c","abstract_canon_sha256":"ed0c704ee768e08c01e45cf1cca1cfb8f4fe811a9c44967a1d95310b86392b79"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:02:13.021277Z","signature_b64":"UVR6blydDgzqvx+gyiwd/vxDeEubX53++8qqcVLKXmNjCo6HUNnPZvG4sny6UsW4gcTHvvk8OAlM6TNUuScCBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5c08b54157216c237e69ef43caeaeb127dce9bcce02d8afd6f4200be312b346c","last_reissued_at":"2026-05-18T00:02:13.020629Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:02:13.020629Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Which Neural Net Architectures Give Rise To Exploding and Vanishing Gradients?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.PR","math.ST","stat.TH"],"primary_cat":"stat.ML","authors_text":"Boris Hanin","submitted_at":"2018-01-11T13:17:22Z","abstract_excerpt":"We give a rigorous analysis of the statistical behavior of gradients in a randomly initialized fully connected network N with ReLU activations. Our results show that the empirical variance of the squares of the entries in the input-output Jacobian of N is exponential in a simple architecture-dependent constant beta, given by the sum of the reciprocals of the hidden layer widths. When beta is large, the gradients computed by N at initialization vary wildly. Our approach complements the mean field theory analysis of random networks. From this point of view, we rigorously compute finite width cor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1801.03744","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":"1801.03744","created_at":"2026-05-18T00:02:13.020734+00:00"},{"alias_kind":"arxiv_version","alias_value":"1801.03744v3","created_at":"2026-05-18T00:02:13.020734+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1801.03744","created_at":"2026-05-18T00:02:13.020734+00:00"},{"alias_kind":"pith_short_12","alias_value":"LQELKQKXEFWC","created_at":"2026-05-18T12:32:37.024351+00:00"},{"alias_kind":"pith_short_16","alias_value":"LQELKQKXEFWCG7TJ","created_at":"2026-05-18T12:32:37.024351+00:00"},{"alias_kind":"pith_short_8","alias_value":"LQELKQKX","created_at":"2026-05-18T12:32:37.024351+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/LQELKQKXEFWCG7TJ55B4V2XLCJ","json":"https://pith.science/pith/LQELKQKXEFWCG7TJ55B4V2XLCJ.json","graph_json":"https://pith.science/api/pith-number/LQELKQKXEFWCG7TJ55B4V2XLCJ/graph.json","events_json":"https://pith.science/api/pith-number/LQELKQKXEFWCG7TJ55B4V2XLCJ/events.json","paper":"https://pith.science/paper/LQELKQKX"},"agent_actions":{"view_html":"https://pith.science/pith/LQELKQKXEFWCG7TJ55B4V2XLCJ","download_json":"https://pith.science/pith/LQELKQKXEFWCG7TJ55B4V2XLCJ.json","view_paper":"https://pith.science/paper/LQELKQKX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1801.03744&json=true","fetch_graph":"https://pith.science/api/pith-number/LQELKQKXEFWCG7TJ55B4V2XLCJ/graph.json","fetch_events":"https://pith.science/api/pith-number/LQELKQKXEFWCG7TJ55B4V2XLCJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LQELKQKXEFWCG7TJ55B4V2XLCJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LQELKQKXEFWCG7TJ55B4V2XLCJ/action/storage_attestation","attest_author":"https://pith.science/pith/LQELKQKXEFWCG7TJ55B4V2XLCJ/action/author_attestation","sign_citation":"https://pith.science/pith/LQELKQKXEFWCG7TJ55B4V2XLCJ/action/citation_signature","submit_replication":"https://pith.science/pith/LQELKQKXEFWCG7TJ55B4V2XLCJ/action/replication_record"}},"created_at":"2026-05-18T00:02:13.020734+00:00","updated_at":"2026-05-18T00:02:13.020734+00:00"}