{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:Z7ITT3HCAANIX6FYQWRLXLVWFT","short_pith_number":"pith:Z7ITT3HC","schema_version":"1.0","canonical_sha256":"cfd139ece2001a8bf8b885a2bbaeb62ce6088ad0c0a26266e91a36c87505bc87","source":{"kind":"arxiv","id":"2106.00903","version":2},"attestation_state":"computed","paper":{"title":"Rejuvenating Low-Frequency Words: Making the Most of Parallel Data in Non-Autoregressive Translation","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Dacheng Tao, Derek F. Wong, Liang Ding, Longyue Wang, Xuebo Liu, Zhaopeng Tu","submitted_at":"2021-06-02T02:41:40Z","abstract_excerpt":"Knowledge distillation (KD) is commonly used to construct synthetic data for training non-autoregressive translation (NAT) models. However, there exists a discrepancy on low-frequency words between the distilled and the original data, leading to more errors on predicting low-frequency words. To alleviate the problem, we directly expose the raw data into NAT by leveraging pretraining. By analyzing directed alignments, we found that KD makes low-frequency source words aligned with targets more deterministically but fails to align sufficient low-frequency words from target to source. Accordingly,"},"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":"2106.00903","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2021-06-02T02:41:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7ae60ee749be86020e9394af62ce77840982999f74ce06104ef4e6bfe05fa555","abstract_canon_sha256":"140bde004e2a4637a06d7144eaac4eee77e7edac44b930fbad7d5ec61030ef42"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:17:35.443113Z","signature_b64":"XAebEyBhDPKwnUbRYC9dicNrmgEvRTUeevXid/2HPs/Bl6/rYPBNgqVM5g7QOLvXHLYvBRXs0PnJ8XElEW0+Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cfd139ece2001a8bf8b885a2bbaeb62ce6088ad0c0a26266e91a36c87505bc87","last_reissued_at":"2026-07-05T04:17:35.442643Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:17:35.442643Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rejuvenating Low-Frequency Words: Making the Most of Parallel Data in Non-Autoregressive Translation","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Dacheng Tao, Derek F. Wong, Liang Ding, Longyue Wang, Xuebo Liu, Zhaopeng Tu","submitted_at":"2021-06-02T02:41:40Z","abstract_excerpt":"Knowledge distillation (KD) is commonly used to construct synthetic data for training non-autoregressive translation (NAT) models. However, there exists a discrepancy on low-frequency words between the distilled and the original data, leading to more errors on predicting low-frequency words. To alleviate the problem, we directly expose the raw data into NAT by leveraging pretraining. By analyzing directed alignments, we found that KD makes low-frequency source words aligned with targets more deterministically but fails to align sufficient low-frequency words from target to source. Accordingly,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.00903","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/2106.00903/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":"2106.00903","created_at":"2026-07-05T04:17:35.442700+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.00903v2","created_at":"2026-07-05T04:17:35.442700+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.00903","created_at":"2026-07-05T04:17:35.442700+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z7ITT3HCAANI","created_at":"2026-07-05T04:17:35.442700+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z7ITT3HCAANIX6FY","created_at":"2026-07-05T04:17:35.442700+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z7ITT3HC","created_at":"2026-07-05T04:17:35.442700+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT","json":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT.json","graph_json":"https://pith.science/api/pith-number/Z7ITT3HCAANIX6FYQWRLXLVWFT/graph.json","events_json":"https://pith.science/api/pith-number/Z7ITT3HCAANIX6FYQWRLXLVWFT/events.json","paper":"https://pith.science/paper/Z7ITT3HC"},"agent_actions":{"view_html":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT","download_json":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT.json","view_paper":"https://pith.science/paper/Z7ITT3HC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.00903&json=true","fetch_graph":"https://pith.science/api/pith-number/Z7ITT3HCAANIX6FYQWRLXLVWFT/graph.json","fetch_events":"https://pith.science/api/pith-number/Z7ITT3HCAANIX6FYQWRLXLVWFT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT/action/storage_attestation","attest_author":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT/action/author_attestation","sign_citation":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT/action/citation_signature","submit_replication":"https://pith.science/pith/Z7ITT3HCAANIX6FYQWRLXLVWFT/action/replication_record"}},"created_at":"2026-07-05T04:17:35.442700+00:00","updated_at":"2026-07-05T04:17:35.442700+00:00"}