{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:5NI6XX4NOFDILMMDJ2WNKKRYUI","short_pith_number":"pith:5NI6XX4N","schema_version":"1.0","canonical_sha256":"eb51ebdf8d714685b1834eacd52a38a21003df6d1771b1ab04195c223f4afd85","source":{"kind":"arxiv","id":"2103.07040","version":1},"attestation_state":"computed","paper":{"title":"Bilingual Dictionary-based Language Model Pretraining for Neural Machine Translation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Haoying Dai, Jiayong Lin, Shuaicheng Zhang, Yusen Lin","submitted_at":"2021-03-12T02:01:22Z","abstract_excerpt":"Recent studies have demonstrated a perceivable improvement on the performance of neural machine translation by applying cross-lingual language model pretraining (Lample and Conneau, 2019), especially the Translation Language Modeling (TLM). To alleviate the need for expensive parallel corpora by TLM, in this work, we incorporate the translation information from dictionaries into the pretraining process and propose a novel Bilingual Dictionary-based Language Model (BDLM). We evaluate our BDLM in Chinese, English, and Romanian. For Chinese-English, we obtained a 55.0 BLEU on WMT-News19 (Tiedeman"},"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":"2103.07040","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-03-12T02:01:22Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"c3ae4177d2af06d3989a256b6f1367dd907c4fda4586452d78d53a633bbf5b75","abstract_canon_sha256":"cb88708d2f326a6f3c7868e01a329022939afb8176cf3f69a2e23cc255ed2133"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:22:27.968016Z","signature_b64":"whaFuHtsJbSSxa6a9MDdwV4FjS72wvs2AddVi+dkOgLufpEQYhSOMqSjAthGQJNzGFFq65DanGM/hWoaHoCoAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eb51ebdf8d714685b1834eacd52a38a21003df6d1771b1ab04195c223f4afd85","last_reissued_at":"2026-07-05T02:22:27.967554Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:22:27.967554Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Bilingual Dictionary-based Language Model Pretraining for Neural Machine Translation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Haoying Dai, Jiayong Lin, Shuaicheng Zhang, Yusen Lin","submitted_at":"2021-03-12T02:01:22Z","abstract_excerpt":"Recent studies have demonstrated a perceivable improvement on the performance of neural machine translation by applying cross-lingual language model pretraining (Lample and Conneau, 2019), especially the Translation Language Modeling (TLM). To alleviate the need for expensive parallel corpora by TLM, in this work, we incorporate the translation information from dictionaries into the pretraining process and propose a novel Bilingual Dictionary-based Language Model (BDLM). We evaluate our BDLM in Chinese, English, and Romanian. For Chinese-English, we obtained a 55.0 BLEU on WMT-News19 (Tiedeman"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.07040","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/2103.07040/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":"2103.07040","created_at":"2026-07-05T02:22:27.967618+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.07040v1","created_at":"2026-07-05T02:22:27.967618+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.07040","created_at":"2026-07-05T02:22:27.967618+00:00"},{"alias_kind":"pith_short_12","alias_value":"5NI6XX4NOFDI","created_at":"2026-07-05T02:22:27.967618+00:00"},{"alias_kind":"pith_short_16","alias_value":"5NI6XX4NOFDILMMD","created_at":"2026-07-05T02:22:27.967618+00:00"},{"alias_kind":"pith_short_8","alias_value":"5NI6XX4N","created_at":"2026-07-05T02:22:27.967618+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.08348","citing_title":"Refining Translations with LLMs: A Constraint-Aware Iterative Prompting Approach","ref_index":19,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI","json":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI.json","graph_json":"https://pith.science/api/pith-number/5NI6XX4NOFDILMMDJ2WNKKRYUI/graph.json","events_json":"https://pith.science/api/pith-number/5NI6XX4NOFDILMMDJ2WNKKRYUI/events.json","paper":"https://pith.science/paper/5NI6XX4N"},"agent_actions":{"view_html":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI","download_json":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI.json","view_paper":"https://pith.science/paper/5NI6XX4N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.07040&json=true","fetch_graph":"https://pith.science/api/pith-number/5NI6XX4NOFDILMMDJ2WNKKRYUI/graph.json","fetch_events":"https://pith.science/api/pith-number/5NI6XX4NOFDILMMDJ2WNKKRYUI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI/action/storage_attestation","attest_author":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI/action/author_attestation","sign_citation":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI/action/citation_signature","submit_replication":"https://pith.science/pith/5NI6XX4NOFDILMMDJ2WNKKRYUI/action/replication_record"}},"created_at":"2026-07-05T02:22:27.967618+00:00","updated_at":"2026-07-05T02:22:27.967618+00:00"}