{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:VKSXEM4VQ5M2OE3BW5UT6JSR4V","short_pith_number":"pith:VKSXEM4V","schema_version":"1.0","canonical_sha256":"aaa57233958759a71361b7693f2651e54bbb32bfb4737a53ad4da86929cdabb2","source":{"kind":"arxiv","id":"1806.05394","version":2},"attestation_state":"computed","paper":{"title":"Dynamical Isometry and a Mean Field Theory of RNNs: Gating Enables Signal Propagation in Recurrent Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Jeffrey Pennington, Minmin Chen, Samuel S. Schoenholz","submitted_at":"2018-06-14T07:04:31Z","abstract_excerpt":"Recurrent neural networks have gained widespread use in modeling sequence data across various domains. While many successful recurrent architectures employ a notion of gating, the exact mechanism that enables such remarkable performance is not well understood. We develop a theory for signal propagation in recurrent networks after random initialization using a combination of mean field theory and random matrix theory. To simplify our discussion, we introduce a new RNN cell with a simple gating mechanism that we call the minimalRNN and compare it with vanilla RNNs. Our theory allows us to define"},"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":"1806.05394","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-06-14T07:04:31Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"c945015d0e70aa35e7cee8674c0c0916bb500fe13c7504df237f24bfd0e5d41a","abstract_canon_sha256":"4f33c3c0870cdc4adc860b21cc9d91b203c415446f5411e0306287bd81c76c97"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:08:02.796378Z","signature_b64":"ORmF2N6b5kRPiAgk1hptsQyUdKkT/8/M0uqkXpwEZnokBuqo0nkiggUufiiERKdx0i2hkqy65lTRN1Rxj2/RDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aaa57233958759a71361b7693f2651e54bbb32bfb4737a53ad4da86929cdabb2","last_reissued_at":"2026-05-18T00:08:02.795655Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:08:02.795655Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dynamical Isometry and a Mean Field Theory of RNNs: Gating Enables Signal Propagation in Recurrent Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Jeffrey Pennington, Minmin Chen, Samuel S. Schoenholz","submitted_at":"2018-06-14T07:04:31Z","abstract_excerpt":"Recurrent neural networks have gained widespread use in modeling sequence data across various domains. While many successful recurrent architectures employ a notion of gating, the exact mechanism that enables such remarkable performance is not well understood. We develop a theory for signal propagation in recurrent networks after random initialization using a combination of mean field theory and random matrix theory. To simplify our discussion, we introduce a new RNN cell with a simple gating mechanism that we call the minimalRNN and compare it with vanilla RNNs. Our theory allows us to define"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1806.05394","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":""},"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":"1806.05394","created_at":"2026-05-18T00:08:02.795766+00:00"},{"alias_kind":"arxiv_version","alias_value":"1806.05394v2","created_at":"2026-05-18T00:08:02.795766+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1806.05394","created_at":"2026-05-18T00:08:02.795766+00:00"},{"alias_kind":"pith_short_12","alias_value":"VKSXEM4VQ5M2","created_at":"2026-05-18T12:32:59.047623+00:00"},{"alias_kind":"pith_short_16","alias_value":"VKSXEM4VQ5M2OE3B","created_at":"2026-05-18T12:32:59.047623+00:00"},{"alias_kind":"pith_short_8","alias_value":"VKSXEM4V","created_at":"2026-05-18T12:32:59.047623+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.09417","citing_title":"Why Post-Norm Transformers Collapse: Attention Amplification and Gradient Repair Failure","ref_index":12,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V","json":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V.json","graph_json":"https://pith.science/api/pith-number/VKSXEM4VQ5M2OE3BW5UT6JSR4V/graph.json","events_json":"https://pith.science/api/pith-number/VKSXEM4VQ5M2OE3BW5UT6JSR4V/events.json","paper":"https://pith.science/paper/VKSXEM4V"},"agent_actions":{"view_html":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V","download_json":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V.json","view_paper":"https://pith.science/paper/VKSXEM4V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1806.05394&json=true","fetch_graph":"https://pith.science/api/pith-number/VKSXEM4VQ5M2OE3BW5UT6JSR4V/graph.json","fetch_events":"https://pith.science/api/pith-number/VKSXEM4VQ5M2OE3BW5UT6JSR4V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V/action/storage_attestation","attest_author":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V/action/author_attestation","sign_citation":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V/action/citation_signature","submit_replication":"https://pith.science/pith/VKSXEM4VQ5M2OE3BW5UT6JSR4V/action/replication_record"}},"created_at":"2026-05-18T00:08:02.795766+00:00","updated_at":"2026-05-18T00:08:02.795766+00:00"}