{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:4RMFAK3LJ5H2B7TYISLB3WHQHW","short_pith_number":"pith:4RMFAK3L","schema_version":"1.0","canonical_sha256":"e458502b6b4f4fa0fe7844961dd8f03d95b31ae1de844a40444b638f38bad280","source":{"kind":"arxiv","id":"1804.04849","version":3},"attestation_state":"computed","paper":{"title":"The unreasonable effectiveness of the forget gate","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.NE","authors_text":"Joan Lasenby, Jos van der Westhuizen","submitted_at":"2018-04-13T09:18:17Z","abstract_excerpt":"Given the success of the gated recurrent unit, a natural question is whether all the gates of the long short-term memory (LSTM) network are necessary. Previous research has shown that the forget gate is one of the most important gates in the LSTM. Here we show that a forget-gate-only version of the LSTM with chrono-initialized biases, not only provides computational savings but outperforms the standard LSTM on multiple benchmark datasets and competes with some of the best contemporary models. Our proposed network, the JANET, achieves accuracies of 99% and 92.5% on the MNIST and pMNIST datasets"},"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":"1804.04849","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.NE","submitted_at":"2018-04-13T09:18:17Z","cross_cats_sorted":["cs.LG","stat.ML"],"title_canon_sha256":"ef4d5125bdf07ad28765e693e8a175c71c837d47b52b88b3698d362dbccff8f2","abstract_canon_sha256":"aaebbac31cd33dfeae66289ee2de4ed008a9df811a58b922a34593eae380cab6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:05:50.822523Z","signature_b64":"NFebrARrhqCctKGl+gr1i5+j6MddQUa0LN95XvENF0Tet5d4V2NumBsMs4/KMeIqFXQHa1L4vttOD7m9wdqqDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e458502b6b4f4fa0fe7844961dd8f03d95b31ae1de844a40444b638f38bad280","last_reissued_at":"2026-05-18T00:05:50.821864Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:05:50.821864Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The unreasonable effectiveness of the forget gate","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"cs.NE","authors_text":"Joan Lasenby, Jos van der Westhuizen","submitted_at":"2018-04-13T09:18:17Z","abstract_excerpt":"Given the success of the gated recurrent unit, a natural question is whether all the gates of the long short-term memory (LSTM) network are necessary. Previous research has shown that the forget gate is one of the most important gates in the LSTM. Here we show that a forget-gate-only version of the LSTM with chrono-initialized biases, not only provides computational savings but outperforms the standard LSTM on multiple benchmark datasets and competes with some of the best contemporary models. Our proposed network, the JANET, achieves accuracies of 99% and 92.5% on the MNIST and pMNIST datasets"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1804.04849","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":"1804.04849","created_at":"2026-05-18T00:05:50.821973+00:00"},{"alias_kind":"arxiv_version","alias_value":"1804.04849v3","created_at":"2026-05-18T00:05:50.821973+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1804.04849","created_at":"2026-05-18T00:05:50.821973+00:00"},{"alias_kind":"pith_short_12","alias_value":"4RMFAK3LJ5H2","created_at":"2026-05-18T12:32:05.422762+00:00"},{"alias_kind":"pith_short_16","alias_value":"4RMFAK3LJ5H2B7TY","created_at":"2026-05-18T12:32:05.422762+00:00"},{"alias_kind":"pith_short_8","alias_value":"4RMFAK3L","created_at":"2026-05-18T12:32:05.422762+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2508.12121","citing_title":"Time-Scale Coupling Between States and Parameters in Recurrent Neural Networks","ref_index":25,"is_internal_anchor":true},{"citing_arxiv_id":"2312.06635","citing_title":"Gated Linear Attention Transformers with Hardware-Efficient Training","ref_index":94,"is_internal_anchor":true},{"citing_arxiv_id":"2501.00663","citing_title":"Titans: Learning to Memorize at Test Time","ref_index":109,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12206","citing_title":"On the Importance of Multistability for Horizon Generalization in Reinforcement Learning","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27063","citing_title":"Learning to Forget: Continual Learning with Adaptive Weight Decay","ref_index":45,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW","json":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW.json","graph_json":"https://pith.science/api/pith-number/4RMFAK3LJ5H2B7TYISLB3WHQHW/graph.json","events_json":"https://pith.science/api/pith-number/4RMFAK3LJ5H2B7TYISLB3WHQHW/events.json","paper":"https://pith.science/paper/4RMFAK3L"},"agent_actions":{"view_html":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW","download_json":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW.json","view_paper":"https://pith.science/paper/4RMFAK3L","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1804.04849&json=true","fetch_graph":"https://pith.science/api/pith-number/4RMFAK3LJ5H2B7TYISLB3WHQHW/graph.json","fetch_events":"https://pith.science/api/pith-number/4RMFAK3LJ5H2B7TYISLB3WHQHW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW/action/storage_attestation","attest_author":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW/action/author_attestation","sign_citation":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW/action/citation_signature","submit_replication":"https://pith.science/pith/4RMFAK3LJ5H2B7TYISLB3WHQHW/action/replication_record"}},"created_at":"2026-05-18T00:05:50.821973+00:00","updated_at":"2026-05-18T00:05:50.821973+00:00"}