{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:SUXILGGR4VCDA4YEGIFBSBY5FI","short_pith_number":"pith:SUXILGGR","schema_version":"1.0","canonical_sha256":"952e8598d1e544307304320a19071d2a1bc31595b40ec047af126505fd0e4023","source":{"kind":"arxiv","id":"1901.11359","version":1},"attestation_state":"computed","paper":{"title":"Adding Interpretable Attention to Neural Translation Models Improves Word Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Joern Wuebker, John DeNero, Thomas Zenkel","submitted_at":"2019-01-31T14:05:02Z","abstract_excerpt":"Multi-layer models with multiple attention heads per layer provide superior translation quality compared to simpler and shallower models, but determining what source context is most relevant to each target word is more challenging as a result. Therefore, deriving high-accuracy word alignments from the activations of a state-of-the-art neural machine translation model is an open challenge. We propose a simple model extension to the Transformer architecture that makes use of its hidden representations and is restricted to attend solely on encoder information to predict the next word. It can be 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":"1901.11359","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-01-31T14:05:02Z","cross_cats_sorted":[],"title_canon_sha256":"c3a4c96ec39d83a263cf9a7b16e3c5735ee833bf163f9299a13d08f21ff1dd9c","abstract_canon_sha256":"49bf57437b055a208d5a77c28aae1240ee84a747ba464ca6ed1cfd5725ade1c9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:55:02.086227Z","signature_b64":"uJpASx6No04WESWamoxKgH3Dk+yUsx5For2hBH0r2oeaUqEvnyTjtwi0he6iNPE+Wa2HgT8+y1NWg9LdPXrVCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"952e8598d1e544307304320a19071d2a1bc31595b40ec047af126505fd0e4023","last_reissued_at":"2026-05-17T23:55:02.085700Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:55:02.085700Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adding Interpretable Attention to Neural Translation Models Improves Word Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Joern Wuebker, John DeNero, Thomas Zenkel","submitted_at":"2019-01-31T14:05:02Z","abstract_excerpt":"Multi-layer models with multiple attention heads per layer provide superior translation quality compared to simpler and shallower models, but determining what source context is most relevant to each target word is more challenging as a result. Therefore, deriving high-accuracy word alignments from the activations of a state-of-the-art neural machine translation model is an open challenge. We propose a simple model extension to the Transformer architecture that makes use of its hidden representations and is restricted to attend solely on encoder information to predict the next word. It can be t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1901.11359","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":"1901.11359","created_at":"2026-05-17T23:55:02.085789+00:00"},{"alias_kind":"arxiv_version","alias_value":"1901.11359v1","created_at":"2026-05-17T23:55:02.085789+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1901.11359","created_at":"2026-05-17T23:55:02.085789+00:00"},{"alias_kind":"pith_short_12","alias_value":"SUXILGGR4VCD","created_at":"2026-05-18T12:33:27.125529+00:00"},{"alias_kind":"pith_short_16","alias_value":"SUXILGGR4VCDA4YE","created_at":"2026-05-18T12:33:27.125529+00:00"},{"alias_kind":"pith_short_8","alias_value":"SUXILGGR","created_at":"2026-05-18T12:33:27.125529+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"1906.10282","citing_title":"Saliency-driven Word Alignment Interpretation for Neural Machine Translation","ref_index":44,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI","json":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI.json","graph_json":"https://pith.science/api/pith-number/SUXILGGR4VCDA4YEGIFBSBY5FI/graph.json","events_json":"https://pith.science/api/pith-number/SUXILGGR4VCDA4YEGIFBSBY5FI/events.json","paper":"https://pith.science/paper/SUXILGGR"},"agent_actions":{"view_html":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI","download_json":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI.json","view_paper":"https://pith.science/paper/SUXILGGR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1901.11359&json=true","fetch_graph":"https://pith.science/api/pith-number/SUXILGGR4VCDA4YEGIFBSBY5FI/graph.json","fetch_events":"https://pith.science/api/pith-number/SUXILGGR4VCDA4YEGIFBSBY5FI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI/action/storage_attestation","attest_author":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI/action/author_attestation","sign_citation":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI/action/citation_signature","submit_replication":"https://pith.science/pith/SUXILGGR4VCDA4YEGIFBSBY5FI/action/replication_record"}},"created_at":"2026-05-17T23:55:02.085789+00:00","updated_at":"2026-05-17T23:55:02.085789+00:00"}