{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:IFF4UWGW7SHXFBW75YBJHLZZGC","short_pith_number":"pith:IFF4UWGW","schema_version":"1.0","canonical_sha256":"414bca58d6fc8f7286dfee0293af39308f158bd15b98c1b16e2e062c3a707cb2","source":{"kind":"arxiv","id":"2606.22478","version":1},"attestation_state":"computed","paper":{"title":"ROMEVA: Geometry-Preserving Vocabulary Expansion for Roman Urdu Language Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Afsheen Asif, Mahnoor Khan, Mehwish Fatima, Milhan Afzal Khan, Seemab Latif","submitted_at":"2026-06-21T12:40:16Z","abstract_excerpt":"Multilingual Language Models like mBERT are widely used for low-resource NLP, yet their adaptation to morphologically inconsistent languages such as Roman Urdu remains underexplored. Roman Urdu spelling variation causes severe sub-word fragmentation, averaging 1.50 sub-words per token. We propose \\textit{ROMEVA} (Roman Urdu Embedding-preserving Vocabulary Adaptation), which combines sub-word-average initialization and a PCA-guided anchor loss to stabilize embeddings during vocabulary expansion. Using a 36,130-comment Roman Urdu corpus, we add 500 highly fragmented tokens to mBERT and compare n"},"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":"2606.22478","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-21T12:40:16Z","cross_cats_sorted":[],"title_canon_sha256":"9a6fb3e4d2f64d5085aecaed7ed04d735bb9fba49e8433a9b54e47e90e60e04c","abstract_canon_sha256":"3e77cc7dd519d675460f44c0fe48a42d0aa3d9979e2f62cc58442ca00aeb9037"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:39.419060Z","signature_b64":"Us24GtKjMX5OziGU8s+wt9e+n23BeQ4xmvR3ppGCsiNz1LhK10b0emg++g5kke323uYSkmr/OOKNKM/3ygFFAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"414bca58d6fc8f7286dfee0293af39308f158bd15b98c1b16e2e062c3a707cb2","last_reissued_at":"2026-06-23T02:13:39.418654Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:39.418654Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ROMEVA: Geometry-Preserving Vocabulary Expansion for Roman Urdu Language Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Afsheen Asif, Mahnoor Khan, Mehwish Fatima, Milhan Afzal Khan, Seemab Latif","submitted_at":"2026-06-21T12:40:16Z","abstract_excerpt":"Multilingual Language Models like mBERT are widely used for low-resource NLP, yet their adaptation to morphologically inconsistent languages such as Roman Urdu remains underexplored. Roman Urdu spelling variation causes severe sub-word fragmentation, averaging 1.50 sub-words per token. We propose \\textit{ROMEVA} (Roman Urdu Embedding-preserving Vocabulary Adaptation), which combines sub-word-average initialization and a PCA-guided anchor loss to stabilize embeddings during vocabulary expansion. Using a 36,130-comment Roman Urdu corpus, we add 500 highly fragmented tokens to mBERT and compare n"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22478","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/2606.22478/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":"2606.22478","created_at":"2026-06-23T02:13:39.418717+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.22478v1","created_at":"2026-06-23T02:13:39.418717+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22478","created_at":"2026-06-23T02:13:39.418717+00:00"},{"alias_kind":"pith_short_12","alias_value":"IFF4UWGW7SHX","created_at":"2026-06-23T02:13:39.418717+00:00"},{"alias_kind":"pith_short_16","alias_value":"IFF4UWGW7SHXFBW7","created_at":"2026-06-23T02:13:39.418717+00:00"},{"alias_kind":"pith_short_8","alias_value":"IFF4UWGW","created_at":"2026-06-23T02:13:39.418717+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC","json":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC.json","graph_json":"https://pith.science/api/pith-number/IFF4UWGW7SHXFBW75YBJHLZZGC/graph.json","events_json":"https://pith.science/api/pith-number/IFF4UWGW7SHXFBW75YBJHLZZGC/events.json","paper":"https://pith.science/paper/IFF4UWGW"},"agent_actions":{"view_html":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC","download_json":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC.json","view_paper":"https://pith.science/paper/IFF4UWGW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.22478&json=true","fetch_graph":"https://pith.science/api/pith-number/IFF4UWGW7SHXFBW75YBJHLZZGC/graph.json","fetch_events":"https://pith.science/api/pith-number/IFF4UWGW7SHXFBW75YBJHLZZGC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC/action/storage_attestation","attest_author":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC/action/author_attestation","sign_citation":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC/action/citation_signature","submit_replication":"https://pith.science/pith/IFF4UWGW7SHXFBW75YBJHLZZGC/action/replication_record"}},"created_at":"2026-06-23T02:13:39.418717+00:00","updated_at":"2026-06-23T02:13:39.418717+00:00"}