{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:XHYZIANEZ6TCNHTL3AFGIPO2AG","short_pith_number":"pith:XHYZIANE","schema_version":"1.0","canonical_sha256":"b9f19401a4cfa6269e6bd80a643dda01b505393f223016fa7e30b700485b97d3","source":{"kind":"arxiv","id":"1912.03334","version":1},"attestation_state":"computed","paper":{"title":"Explaining Sequence-Level Knowledge Distillation as Data-Augmentation for Neural Machine Translation","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kevin Duh, Mitchell A. Gordon","submitted_at":"2019-12-06T20:27:38Z","abstract_excerpt":"Sequence-level knowledge distillation (SLKD) is a model compression technique that leverages large, accurate teacher models to train smaller, under-parameterized student models. Why does pre-processing MT data with SLKD help us train smaller models? We test the common hypothesis that SLKD addresses a capacity deficiency in students by \"simplifying\" noisy data points and find it unlikely in our case. Models trained on concatenations of original and \"simplified\" datasets generalize just as well as baseline SLKD. We then propose an alternative hypothesis under the lens of data augmentation and re"},"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":"1912.03334","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2019-12-06T20:27:38Z","cross_cats_sorted":[],"title_canon_sha256":"dc56dcd1dd2289c50da69c1bdba777a1a6ed64d69d0da4d81a784db5da334c39","abstract_canon_sha256":"63e2c8f58c9b85e1fa5913b6b33b2e1c28da21c56e89d0d007d2435ca95732ce"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:24:44.536168Z","signature_b64":"gfpyN7S8TxYCQe+UtOadjz/1GcH9oN0opAIqD9/wKYHctzdLbXn2/AlBU4Ciz8X50Fqd/CRfelq8AkVgwd1DDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b9f19401a4cfa6269e6bd80a643dda01b505393f223016fa7e30b700485b97d3","last_reissued_at":"2026-07-05T00:24:44.535702Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:24:44.535702Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Explaining Sequence-Level Knowledge Distillation as Data-Augmentation for Neural Machine Translation","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kevin Duh, Mitchell A. Gordon","submitted_at":"2019-12-06T20:27:38Z","abstract_excerpt":"Sequence-level knowledge distillation (SLKD) is a model compression technique that leverages large, accurate teacher models to train smaller, under-parameterized student models. Why does pre-processing MT data with SLKD help us train smaller models? We test the common hypothesis that SLKD addresses a capacity deficiency in students by \"simplifying\" noisy data points and find it unlikely in our case. Models trained on concatenations of original and \"simplified\" datasets generalize just as well as baseline SLKD. We then propose an alternative hypothesis under the lens of data augmentation and re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1912.03334","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1912.03334/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":"1912.03334","created_at":"2026-07-05T00:24:44.535762+00:00"},{"alias_kind":"arxiv_version","alias_value":"1912.03334v1","created_at":"2026-07-05T00:24:44.535762+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1912.03334","created_at":"2026-07-05T00:24:44.535762+00:00"},{"alias_kind":"pith_short_12","alias_value":"XHYZIANEZ6TC","created_at":"2026-07-05T00:24:44.535762+00:00"},{"alias_kind":"pith_short_16","alias_value":"XHYZIANEZ6TCNHTL","created_at":"2026-07-05T00:24:44.535762+00:00"},{"alias_kind":"pith_short_8","alias_value":"XHYZIANE","created_at":"2026-07-05T00:24:44.535762+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.01491","citing_title":"Memorization Inheritance in Sequence-Level Knowledge Distillation for Neural Machine Translation","ref_index":4,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG","json":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG.json","graph_json":"https://pith.science/api/pith-number/XHYZIANEZ6TCNHTL3AFGIPO2AG/graph.json","events_json":"https://pith.science/api/pith-number/XHYZIANEZ6TCNHTL3AFGIPO2AG/events.json","paper":"https://pith.science/paper/XHYZIANE"},"agent_actions":{"view_html":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG","download_json":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG.json","view_paper":"https://pith.science/paper/XHYZIANE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1912.03334&json=true","fetch_graph":"https://pith.science/api/pith-number/XHYZIANEZ6TCNHTL3AFGIPO2AG/graph.json","fetch_events":"https://pith.science/api/pith-number/XHYZIANEZ6TCNHTL3AFGIPO2AG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG/action/storage_attestation","attest_author":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG/action/author_attestation","sign_citation":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG/action/citation_signature","submit_replication":"https://pith.science/pith/XHYZIANEZ6TCNHTL3AFGIPO2AG/action/replication_record"}},"created_at":"2026-07-05T00:24:44.535762+00:00","updated_at":"2026-07-05T00:24:44.535762+00:00"}