{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:YH4RGUPKH6Z4BBBYPPWFSNQZXV","short_pith_number":"pith:YH4RGUPK","schema_version":"1.0","canonical_sha256":"c1f91351ea3fb3c084387bec593619bd49271a6a0cf08b9113d7a9d6098a60fe","source":{"kind":"arxiv","id":"2001.04589","version":2},"attestation_state":"computed","paper":{"title":"Faster Transformer Decoding: N-gram Masked Self-Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ankur Bapna, Ciprian Chelba, Mia Chen, Noam Shazeer","submitted_at":"2020-01-14T02:14:09Z","abstract_excerpt":"Motivated by the fact that most of the information relevant to the prediction of target tokens is drawn from the source sentence $S=s_1, \\ldots, s_S$, we propose truncating the target-side window used for computing self-attention by making an $N$-gram assumption. Experiments on WMT EnDe and EnFr data sets show that the $N$-gram masked self-attention model loses very little in BLEU score for $N$ values in the range $4, \\ldots, 8$, depending on the task."},"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":"2001.04589","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-01-14T02:14:09Z","cross_cats_sorted":["cs.CL","stat.ML"],"title_canon_sha256":"ca0aa808105c1720b0880c41450bc4aa30da7f9852d0237096ed34e46373d9fb","abstract_canon_sha256":"4ec2ca85cf39c5353029b147116db3d9691cdae8488cbe2b310f7db7fa297b2d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:50:41.741233Z","signature_b64":"FMaXQgRVC8FbVqj/Qnc/bxfGDIEP21JjcnZjWyWo8AwPub2KytZvJCEkqzCleYyAp4MlEBdIj9bgNiQCNNWgAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c1f91351ea3fb3c084387bec593619bd49271a6a0cf08b9113d7a9d6098a60fe","last_reissued_at":"2026-07-05T09:50:41.740698Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:50:41.740698Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Faster Transformer Decoding: N-gram Masked Self-Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","stat.ML"],"primary_cat":"cs.LG","authors_text":"Ankur Bapna, Ciprian Chelba, Mia Chen, Noam Shazeer","submitted_at":"2020-01-14T02:14:09Z","abstract_excerpt":"Motivated by the fact that most of the information relevant to the prediction of target tokens is drawn from the source sentence $S=s_1, \\ldots, s_S$, we propose truncating the target-side window used for computing self-attention by making an $N$-gram assumption. Experiments on WMT EnDe and EnFr data sets show that the $N$-gram masked self-attention model loses very little in BLEU score for $N$ values in the range $4, \\ldots, 8$, depending on the task."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2001.04589","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2001.04589/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"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":"2001.04589","created_at":"2026-07-05T09:50:41.740769+00:00"},{"alias_kind":"arxiv_version","alias_value":"2001.04589v2","created_at":"2026-07-05T09:50:41.740769+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2001.04589","created_at":"2026-07-05T09:50:41.740769+00:00"},{"alias_kind":"pith_short_12","alias_value":"YH4RGUPKH6Z4","created_at":"2026-07-05T09:50:41.740769+00:00"},{"alias_kind":"pith_short_16","alias_value":"YH4RGUPKH6Z4BBBY","created_at":"2026-07-05T09:50:41.740769+00:00"},{"alias_kind":"pith_short_8","alias_value":"YH4RGUPK","created_at":"2026-07-05T09:50:41.740769+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2009.14794","citing_title":"Rethinking Attention with Performers","ref_index":110,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV","json":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV.json","graph_json":"https://pith.science/api/pith-number/YH4RGUPKH6Z4BBBYPPWFSNQZXV/graph.json","events_json":"https://pith.science/api/pith-number/YH4RGUPKH6Z4BBBYPPWFSNQZXV/events.json","paper":"https://pith.science/paper/YH4RGUPK"},"agent_actions":{"view_html":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV","download_json":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV.json","view_paper":"https://pith.science/paper/YH4RGUPK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2001.04589&json=true","fetch_graph":"https://pith.science/api/pith-number/YH4RGUPKH6Z4BBBYPPWFSNQZXV/graph.json","fetch_events":"https://pith.science/api/pith-number/YH4RGUPKH6Z4BBBYPPWFSNQZXV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV/action/storage_attestation","attest_author":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV/action/author_attestation","sign_citation":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV/action/citation_signature","submit_replication":"https://pith.science/pith/YH4RGUPKH6Z4BBBYPPWFSNQZXV/action/replication_record"}},"created_at":"2026-07-05T09:50:41.740769+00:00","updated_at":"2026-07-05T09:50:41.740769+00:00"}