{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:IDB3LIOTP2XJTTZORVGYSDXQKR","short_pith_number":"pith:IDB3LIOT","schema_version":"1.0","canonical_sha256":"40c3b5a1d37eae99cf2e8d4d890ef05460ab789a3faad84464b3e86375251991","source":{"kind":"arxiv","id":"1908.05204","version":2},"attestation_state":"computed","paper":{"title":"On The Evaluation of Machine Translation Systems Trained With Back-Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Marc'Aurelio Ranzato, Michael Auli, Myle Ott, Sergey Edunov","submitted_at":"2019-08-14T16:24:56Z","abstract_excerpt":"Back-translation is a widely used data augmentation technique which leverages target monolingual data. However, its effectiveness has been challenged since automatic metrics such as BLEU only show significant improvements for test examples where the source itself is a translation, or translationese. This is believed to be due to translationese inputs better matching the back-translated training data. In this work, we show that this conjecture is not empirically supported and that back-translation improves translation quality of both naturally occurring text as well as translationese according "},"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":"1908.05204","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-08-14T16:24:56Z","cross_cats_sorted":[],"title_canon_sha256":"e57bcd1439e2dd2f94ba4e7ff1c1876dd4ef7886d0a505e9f8694bdf25daca6a","abstract_canon_sha256":"1692785bf7ab35e0d24e30b25f2b1d949a7991e3c4cf7b31d376624e2da301f7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:27:43.202778Z","signature_b64":"FgUt+Bli6D6TGRm8uIuoa4/MGta89VkARqPoF9YzcMUreEGhXYONX7aGE3CNSR3ftIhdW/1uiWGSJ3g3rbIODA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"40c3b5a1d37eae99cf2e8d4d890ef05460ab789a3faad84464b3e86375251991","last_reissued_at":"2026-07-05T01:27:43.202364Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:27:43.202364Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On The Evaluation of Machine Translation Systems Trained With Back-Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Marc'Aurelio Ranzato, Michael Auli, Myle Ott, Sergey Edunov","submitted_at":"2019-08-14T16:24:56Z","abstract_excerpt":"Back-translation is a widely used data augmentation technique which leverages target monolingual data. However, its effectiveness has been challenged since automatic metrics such as BLEU only show significant improvements for test examples where the source itself is a translation, or translationese. This is believed to be due to translationese inputs better matching the back-translated training data. In this work, we show that this conjecture is not empirically supported and that back-translation improves translation quality of both naturally occurring text as well as translationese according "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1908.05204","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/1908.05204/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":"1908.05204","created_at":"2026-07-05T01:27:43.202427+00:00"},{"alias_kind":"arxiv_version","alias_value":"1908.05204v2","created_at":"2026-07-05T01:27:43.202427+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1908.05204","created_at":"2026-07-05T01:27:43.202427+00:00"},{"alias_kind":"pith_short_12","alias_value":"IDB3LIOTP2XJ","created_at":"2026-07-05T01:27:43.202427+00:00"},{"alias_kind":"pith_short_16","alias_value":"IDB3LIOTP2XJTTZO","created_at":"2026-07-05T01:27:43.202427+00:00"},{"alias_kind":"pith_short_8","alias_value":"IDB3LIOT","created_at":"2026-07-05T01:27:43.202427+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2504.18413","citing_title":"An Empirical Study of Evaluating Long-form Question Answering","ref_index":10,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR","json":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR.json","graph_json":"https://pith.science/api/pith-number/IDB3LIOTP2XJTTZORVGYSDXQKR/graph.json","events_json":"https://pith.science/api/pith-number/IDB3LIOTP2XJTTZORVGYSDXQKR/events.json","paper":"https://pith.science/paper/IDB3LIOT"},"agent_actions":{"view_html":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR","download_json":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR.json","view_paper":"https://pith.science/paper/IDB3LIOT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1908.05204&json=true","fetch_graph":"https://pith.science/api/pith-number/IDB3LIOTP2XJTTZORVGYSDXQKR/graph.json","fetch_events":"https://pith.science/api/pith-number/IDB3LIOTP2XJTTZORVGYSDXQKR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR/action/storage_attestation","attest_author":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR/action/author_attestation","sign_citation":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR/action/citation_signature","submit_replication":"https://pith.science/pith/IDB3LIOTP2XJTTZORVGYSDXQKR/action/replication_record"}},"created_at":"2026-07-05T01:27:43.202427+00:00","updated_at":"2026-07-05T01:27:43.202427+00:00"}