{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:OIDFK2EEDEQQ5DFIJXCT3PODFU","short_pith_number":"pith:OIDFK2EE","schema_version":"1.0","canonical_sha256":"720655688419210e8ca84dc53dbdc32d3391d03bfc0b7ff251ee904376dc5bb4","source":{"kind":"arxiv","id":"2010.14571","version":2},"attestation_state":"computed","paper":{"title":"Language ID in the Wild: Unexpected Challenges on the Path to a Thousand-Language Web Text Corpus","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Ankur Bapna, Daan van Esch, Isaac Caswell, Theresa Breiner","submitted_at":"2020-10-27T19:29:17Z","abstract_excerpt":"Large text corpora are increasingly important for a wide variety of Natural Language Processing (NLP) tasks, and automatic language identification (LangID) is a core technology needed to collect such datasets in a multilingual context. LangID is largely treated as solved in the literature, with models reported that achieve over 90% average F1 on as many as 1,366 languages. We train LangID models on up to 1,629 languages with comparable quality on held-out test sets, but find that human-judged LangID accuracy for web-crawl text corpora created using these models is only around 5% for many lower"},"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":"2010.14571","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2020-10-27T19:29:17Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8a9f1da9ffa8ae0d21d90477d143754de5510cce81bc6e95e68f2e871c9a91bd","abstract_canon_sha256":"5f5611f0b0e8474feacb968a97b31ef1e70e60c63182f864f9cc0e908e815a4f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:47:33.717778Z","signature_b64":"xI+fPmECVmd7kobMxM3vN9OC/0oEc5j6L1wUxSlOto3xsqhj5dcc03Ihpv7Y4d+sCgjshC7AKwQ475C+9qTPAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"720655688419210e8ca84dc53dbdc32d3391d03bfc0b7ff251ee904376dc5bb4","last_reissued_at":"2026-07-05T01:47:33.717228Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:47:33.717228Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language ID in the Wild: Unexpected Challenges on the Path to a Thousand-Language Web Text Corpus","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Ankur Bapna, Daan van Esch, Isaac Caswell, Theresa Breiner","submitted_at":"2020-10-27T19:29:17Z","abstract_excerpt":"Large text corpora are increasingly important for a wide variety of Natural Language Processing (NLP) tasks, and automatic language identification (LangID) is a core technology needed to collect such datasets in a multilingual context. LangID is largely treated as solved in the literature, with models reported that achieve over 90% average F1 on as many as 1,366 languages. We train LangID models on up to 1,629 languages with comparable quality on held-out test sets, but find that human-judged LangID accuracy for web-crawl text corpora created using these models is only around 5% for many lower"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.14571","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/2010.14571/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":"2010.14571","created_at":"2026-07-05T01:47:33.717311+00:00"},{"alias_kind":"arxiv_version","alias_value":"2010.14571v2","created_at":"2026-07-05T01:47:33.717311+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.14571","created_at":"2026-07-05T01:47:33.717311+00:00"},{"alias_kind":"pith_short_12","alias_value":"OIDFK2EEDEQQ","created_at":"2026-07-05T01:47:33.717311+00:00"},{"alias_kind":"pith_short_16","alias_value":"OIDFK2EEDEQQ5DFI","created_at":"2026-07-05T01:47:33.717311+00:00"},{"alias_kind":"pith_short_8","alias_value":"OIDFK2EE","created_at":"2026-07-05T01:47:33.717311+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2101.00027","citing_title":"The Pile: An 800GB Dataset of Diverse Text for Language Modeling","ref_index":183,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU","json":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU.json","graph_json":"https://pith.science/api/pith-number/OIDFK2EEDEQQ5DFIJXCT3PODFU/graph.json","events_json":"https://pith.science/api/pith-number/OIDFK2EEDEQQ5DFIJXCT3PODFU/events.json","paper":"https://pith.science/paper/OIDFK2EE"},"agent_actions":{"view_html":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU","download_json":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU.json","view_paper":"https://pith.science/paper/OIDFK2EE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2010.14571&json=true","fetch_graph":"https://pith.science/api/pith-number/OIDFK2EEDEQQ5DFIJXCT3PODFU/graph.json","fetch_events":"https://pith.science/api/pith-number/OIDFK2EEDEQQ5DFIJXCT3PODFU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU/action/storage_attestation","attest_author":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU/action/author_attestation","sign_citation":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU/action/citation_signature","submit_replication":"https://pith.science/pith/OIDFK2EEDEQQ5DFIJXCT3PODFU/action/replication_record"}},"created_at":"2026-07-05T01:47:33.717311+00:00","updated_at":"2026-07-05T01:47:33.717311+00:00"}