{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:EOBEJWAH6Q6WG2XD7HWQHMIUH4","short_pith_number":"pith:EOBEJWAH","schema_version":"1.0","canonical_sha256":"238244d807f43d636ae3f9ed03b1143f27a90f8407d9d033b903e7b40c49159d","source":{"kind":"arxiv","id":"1805.03716","version":1},"attestation_state":"computed","paper":{"title":"Long Short-Term Memory as a Dynamically Computed Element-wise Weighted Sum","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Kenton Lee, Luke Zettlemoyer, Nicholas FitzGerald, Omer Levy","submitted_at":"2018-05-09T20:05:58Z","abstract_excerpt":"LSTMs were introduced to combat vanishing gradients in simple RNNs by augmenting them with gated additive recurrent connections. We present an alternative view to explain the success of LSTMs: the gates themselves are versatile recurrent models that provide more representational power than previously appreciated. We do this by decoupling the LSTM's gates from the embedded simple RNN, producing a new class of RNNs where the recurrence computes an element-wise weighted sum of context-independent functions of the input. Ablations on a range of problems demonstrate that the gating mechanism alone "},"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":"1805.03716","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2018-05-09T20:05:58Z","cross_cats_sorted":["cs.AI","cs.LG","stat.ML"],"title_canon_sha256":"00770d500b057e071c470ef91246a4b23aa6e8bb660c0781503de9c55cfdf0c6","abstract_canon_sha256":"4dafdb95239cf8fae1a361ad4563b31684aede4a9d37d0da7b67bc350ec10ee8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:16:18.944264Z","signature_b64":"WyIiavsqOHEmF891R70dxIc33Ovg/gmcuBhbUZD9DB7TIaad6gMwotQDOz/3XQzyHwhHvKTjpc8ndJLZKxwOAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"238244d807f43d636ae3f9ed03b1143f27a90f8407d9d033b903e7b40c49159d","last_reissued_at":"2026-05-18T00:16:18.943719Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:16:18.943719Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Long Short-Term Memory as a Dynamically Computed Element-wise Weighted Sum","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.ML"],"primary_cat":"cs.CL","authors_text":"Kenton Lee, Luke Zettlemoyer, Nicholas FitzGerald, Omer Levy","submitted_at":"2018-05-09T20:05:58Z","abstract_excerpt":"LSTMs were introduced to combat vanishing gradients in simple RNNs by augmenting them with gated additive recurrent connections. We present an alternative view to explain the success of LSTMs: the gates themselves are versatile recurrent models that provide more representational power than previously appreciated. We do this by decoupling the LSTM's gates from the embedded simple RNN, producing a new class of RNNs where the recurrence computes an element-wise weighted sum of context-independent functions of the input. Ablations on a range of problems demonstrate that the gating mechanism alone "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.03716","kind":"arxiv","version":1},"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":"1805.03716","created_at":"2026-05-18T00:16:18.943802+00:00"},{"alias_kind":"arxiv_version","alias_value":"1805.03716v1","created_at":"2026-05-18T00:16:18.943802+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.03716","created_at":"2026-05-18T00:16:18.943802+00:00"},{"alias_kind":"pith_short_12","alias_value":"EOBEJWAH6Q6W","created_at":"2026-05-18T12:32:22.470017+00:00"},{"alias_kind":"pith_short_16","alias_value":"EOBEJWAH6Q6WG2XD","created_at":"2026-05-18T12:32:22.470017+00:00"},{"alias_kind":"pith_short_8","alias_value":"EOBEJWAH","created_at":"2026-05-18T12:32:22.470017+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"1908.00588","citing_title":"Visualizing RNN States with Predictive Semantic Encodings","ref_index":7,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4","json":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4.json","graph_json":"https://pith.science/api/pith-number/EOBEJWAH6Q6WG2XD7HWQHMIUH4/graph.json","events_json":"https://pith.science/api/pith-number/EOBEJWAH6Q6WG2XD7HWQHMIUH4/events.json","paper":"https://pith.science/paper/EOBEJWAH"},"agent_actions":{"view_html":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4","download_json":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4.json","view_paper":"https://pith.science/paper/EOBEJWAH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1805.03716&json=true","fetch_graph":"https://pith.science/api/pith-number/EOBEJWAH6Q6WG2XD7HWQHMIUH4/graph.json","fetch_events":"https://pith.science/api/pith-number/EOBEJWAH6Q6WG2XD7HWQHMIUH4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4/action/storage_attestation","attest_author":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4/action/author_attestation","sign_citation":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4/action/citation_signature","submit_replication":"https://pith.science/pith/EOBEJWAH6Q6WG2XD7HWQHMIUH4/action/replication_record"}},"created_at":"2026-05-18T00:16:18.943802+00:00","updated_at":"2026-05-18T00:16:18.943802+00:00"}