{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:M3HDJVKSUYEPBHG7TS4E6D23TE","short_pith_number":"pith:M3HDJVKS","schema_version":"1.0","canonical_sha256":"66ce34d552a608f09cdf9cb84f0f5b9933f90a90c15f11090e444e82e3d81532","source":{"kind":"arxiv","id":"1911.02727","version":3},"attestation_state":"computed","paper":{"title":"Understanding Knowledge Distillation in Non-autoregressive Machine Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chunting Zhou, Graham Neubig, Jiatao Gu","submitted_at":"2019-11-07T02:47:26Z","abstract_excerpt":"Non-autoregressive machine translation (NAT) systems predict a sequence of output tokens in parallel, achieving substantial improvements in generation speed compared to autoregressive models. Existing NAT models usually rely on the technique of knowledge distillation, which creates the training data from a pretrained autoregressive model for better performance. Knowledge distillation is empirically useful, leading to large gains in accuracy for NAT models, but the reason for this success has, as of yet, been unclear. In this paper, we first design systematic experiments to investigate why know"},"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":"1911.02727","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-11-07T02:47:26Z","cross_cats_sorted":[],"title_canon_sha256":"fce7fdd6aab7c08764440b0fc64a5a038f047ba4f7415be94c0d0d0229e52071","abstract_canon_sha256":"d598e5b444e437c0e5848a1931cdb47051792c660f3672a4fd4c420bb109c8d8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:17:07.554905Z","signature_b64":"/8Fuokc5au+tUnbFoA79MGnPfo1LgkPtgWOrTKN2rNdfqkjhCYvCeknnAoEaRQaw0JzKoOe4A8wm8B2W0YqgBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"66ce34d552a608f09cdf9cb84f0f5b9933f90a90c15f11090e444e82e3d81532","last_reissued_at":"2026-07-05T02:17:07.554430Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:17:07.554430Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding Knowledge Distillation in Non-autoregressive Machine Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chunting Zhou, Graham Neubig, Jiatao Gu","submitted_at":"2019-11-07T02:47:26Z","abstract_excerpt":"Non-autoregressive machine translation (NAT) systems predict a sequence of output tokens in parallel, achieving substantial improvements in generation speed compared to autoregressive models. Existing NAT models usually rely on the technique of knowledge distillation, which creates the training data from a pretrained autoregressive model for better performance. Knowledge distillation is empirically useful, leading to large gains in accuracy for NAT models, but the reason for this success has, as of yet, been unclear. In this paper, we first design systematic experiments to investigate why know"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1911.02727","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1911.02727/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":"1911.02727","created_at":"2026-07-05T02:17:07.554494+00:00"},{"alias_kind":"arxiv_version","alias_value":"1911.02727v3","created_at":"2026-07-05T02:17:07.554494+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1911.02727","created_at":"2026-07-05T02:17:07.554494+00:00"},{"alias_kind":"pith_short_12","alias_value":"M3HDJVKSUYEP","created_at":"2026-07-05T02:17:07.554494+00:00"},{"alias_kind":"pith_short_16","alias_value":"M3HDJVKSUYEPBHG7","created_at":"2026-07-05T02:17:07.554494+00:00"},{"alias_kind":"pith_short_8","alias_value":"M3HDJVKS","created_at":"2026-07-05T02:17:07.554494+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.09924","citing_title":"Evolving Knowledge Distillation for Lightweight Neural Machine Translation","ref_index":35,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE","json":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE.json","graph_json":"https://pith.science/api/pith-number/M3HDJVKSUYEPBHG7TS4E6D23TE/graph.json","events_json":"https://pith.science/api/pith-number/M3HDJVKSUYEPBHG7TS4E6D23TE/events.json","paper":"https://pith.science/paper/M3HDJVKS"},"agent_actions":{"view_html":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE","download_json":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE.json","view_paper":"https://pith.science/paper/M3HDJVKS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1911.02727&json=true","fetch_graph":"https://pith.science/api/pith-number/M3HDJVKSUYEPBHG7TS4E6D23TE/graph.json","fetch_events":"https://pith.science/api/pith-number/M3HDJVKSUYEPBHG7TS4E6D23TE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE/action/storage_attestation","attest_author":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE/action/author_attestation","sign_citation":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE/action/citation_signature","submit_replication":"https://pith.science/pith/M3HDJVKSUYEPBHG7TS4E6D23TE/action/replication_record"}},"created_at":"2026-07-05T02:17:07.554494+00:00","updated_at":"2026-07-05T02:17:07.554494+00:00"}