{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RAJXCWS2CJHRPYC3RCOY4LHBDG","short_pith_number":"pith:RAJXCWS2","schema_version":"1.0","canonical_sha256":"8813715a5a124f17e05b889d8e2ce119a413006d17173b632c63cf8be4d46580","source":{"kind":"arxiv","id":"2405.12413","version":1},"attestation_state":"computed","paper":{"title":"Targeted Multilingual Adaptation for Low-resource Language Families","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"C.M. Downey, Dhwani Serai, Dwija Parikh, Shane Steinert-Threlkeld, Terra Blevins","submitted_at":"2024-05-20T23:38:06Z","abstract_excerpt":"The \"massively-multilingual\" training of multilingual models is known to limit their utility in any one language, and they perform particularly poorly on low-resource languages. However, there is evidence that low-resource languages can benefit from targeted multilinguality, where the model is trained on closely related languages. To test this approach more rigorously, we systematically study best practices for adapting a pre-trained model to a language family. Focusing on the Uralic family as a test case, we adapt XLM-R under various configurations to model 15 languages; we then evaluate the "},"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":"2405.12413","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-20T23:38:06Z","cross_cats_sorted":[],"title_canon_sha256":"bd1cb8dbf21e9751552ad1a1cd5f9bbc4631777f0f80c5fc07d37dd846f87683","abstract_canon_sha256":"3eb0924efd9c3b1ad476541caac41b14684de01a5024858fb488fd4cfa495d70"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:21:19.691530Z","signature_b64":"W7vVgSPAqOpPcejJxtfZbBvDCm18PyoXOVJkE2DxF1vFlgPQ+nlHlCkopboskWLW3qAIRE6tsJgWsZj0ruRfCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8813715a5a124f17e05b889d8e2ce119a413006d17173b632c63cf8be4d46580","last_reissued_at":"2026-07-05T08:21:19.691059Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:21:19.691059Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Targeted Multilingual Adaptation for Low-resource Language Families","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"C.M. Downey, Dhwani Serai, Dwija Parikh, Shane Steinert-Threlkeld, Terra Blevins","submitted_at":"2024-05-20T23:38:06Z","abstract_excerpt":"The \"massively-multilingual\" training of multilingual models is known to limit their utility in any one language, and they perform particularly poorly on low-resource languages. However, there is evidence that low-resource languages can benefit from targeted multilinguality, where the model is trained on closely related languages. To test this approach more rigorously, we systematically study best practices for adapting a pre-trained model to a language family. Focusing on the Uralic family as a test case, we adapt XLM-R under various configurations to model 15 languages; we then evaluate the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.12413","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/2405.12413/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":"2405.12413","created_at":"2026-07-05T08:21:19.691120+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.12413v1","created_at":"2026-07-05T08:21:19.691120+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.12413","created_at":"2026-07-05T08:21:19.691120+00:00"},{"alias_kind":"pith_short_12","alias_value":"RAJXCWS2CJHR","created_at":"2026-07-05T08:21:19.691120+00:00"},{"alias_kind":"pith_short_16","alias_value":"RAJXCWS2CJHRPYC3","created_at":"2026-07-05T08:21:19.691120+00:00"},{"alias_kind":"pith_short_8","alias_value":"RAJXCWS2","created_at":"2026-07-05T08:21:19.691120+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.23394","citing_title":"Teaching a Language Model to Speak the Language of Tools","ref_index":25,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG","json":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG.json","graph_json":"https://pith.science/api/pith-number/RAJXCWS2CJHRPYC3RCOY4LHBDG/graph.json","events_json":"https://pith.science/api/pith-number/RAJXCWS2CJHRPYC3RCOY4LHBDG/events.json","paper":"https://pith.science/paper/RAJXCWS2"},"agent_actions":{"view_html":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG","download_json":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG.json","view_paper":"https://pith.science/paper/RAJXCWS2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.12413&json=true","fetch_graph":"https://pith.science/api/pith-number/RAJXCWS2CJHRPYC3RCOY4LHBDG/graph.json","fetch_events":"https://pith.science/api/pith-number/RAJXCWS2CJHRPYC3RCOY4LHBDG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG/action/storage_attestation","attest_author":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG/action/author_attestation","sign_citation":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG/action/citation_signature","submit_replication":"https://pith.science/pith/RAJXCWS2CJHRPYC3RCOY4LHBDG/action/replication_record"}},"created_at":"2026-07-05T08:21:19.691120+00:00","updated_at":"2026-07-05T08:21:19.691120+00:00"}