{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:OV4SSAHPIPZCVVMNDPQR7ZCQB4","short_pith_number":"pith:OV4SSAHP","schema_version":"1.0","canonical_sha256":"75792900ef43f22ad58d1be11fe4500f304138f1b50c94237e0dcbc974d9c60d","source":{"kind":"arxiv","id":"2605.15854","version":1},"attestation_state":"computed","paper":{"title":"Improving Automatic Speech Recognition for Speakers Treated for Oral Cancer using Data Augmentation and LLM Error Correction","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Combining data augmentation and LLM error correction cuts word error rates by 40-50% for oral cancer speech recognition.","cross_cats":[],"primary_cat":"eess.AS","authors_text":"Bence Mark Halpern, Hidde Folkertsma, Jiapan Guo, Max Witjes, Rob van Son, Sebastiaan de Visscher, Thomas Tienkamp","submitted_at":"2026-05-15T11:13:25Z","abstract_excerpt":"In recent years, the performance of automatic speech recognition (ASR) systems has made considerable progress. Unfortunately, for people with speech impairments, such as people treated for oral cancer (OC), ASR performance is still lagging behind. The scarcity and variability of OC speech data makes development of ASR models for this type of speech difficult. In this work, we use data augmentation and large language model (LLM) error correction to mitigate this problem. We apply various augmentation techniques on a corpus of Dutch oral cancer speech to create synthetic data, and evaluate their"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.15854","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.AS","submitted_at":"2026-05-15T11:13:25Z","cross_cats_sorted":[],"title_canon_sha256":"97dbdc4f0101ffe7ae96c8cc0b137102f5762d08b8a8445ad45215b0e78ff8d6","abstract_canon_sha256":"0d49769322e83b9d2455fc223f7be6c52ad783ee4640cdf2c10f840f962254a2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:22.054711Z","signature_b64":"5VVy2UC60bmVzanXC3XvlJZTjrjftnZ2Nw70umV3ycQNIorX9/W4E0ybi/1JQKeExz2ryG2Hhu0DxOyq0GMdBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"75792900ef43f22ad58d1be11fe4500f304138f1b50c94237e0dcbc974d9c60d","last_reissued_at":"2026-05-20T00:01:22.053845Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:22.053845Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Automatic Speech Recognition for Speakers Treated for Oral Cancer using Data Augmentation and LLM Error Correction","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Combining data augmentation and LLM error correction cuts word error rates by 40-50% for oral cancer speech recognition.","cross_cats":[],"primary_cat":"eess.AS","authors_text":"Bence Mark Halpern, Hidde Folkertsma, Jiapan Guo, Max Witjes, Rob van Son, Sebastiaan de Visscher, Thomas Tienkamp","submitted_at":"2026-05-15T11:13:25Z","abstract_excerpt":"In recent years, the performance of automatic speech recognition (ASR) systems has made considerable progress. Unfortunately, for people with speech impairments, such as people treated for oral cancer (OC), ASR performance is still lagging behind. The scarcity and variability of OC speech data makes development of ASR models for this type of speech difficult. In this work, we use data augmentation and large language model (LLM) error correction to mitigate this problem. We apply various augmentation techniques on a corpus of Dutch oral cancer speech to create synthetic data, and evaluate their"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Overall, we achieve a 40% relative WER decrease for Whisper and a 50% relative WER decrease for MMS, indicating that a combination of data augmentation and LLM correction is a viable strategy for the recognition of OC speech.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The synthetic data produced by the augmentation techniques sufficiently captures the acoustic variability of real oral cancer speech and the LLM corrections do not systematically alter medically relevant content.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TTS data augmentation and LLM error correction together cut relative WER by 40-50% on ASR models for oral cancer speech.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Combining data augmentation and LLM error correction cuts word error rates by 40-50% for oral cancer speech recognition.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"8b75e79284669a48f124dfc314e5f349c369fc6149fbc9e99d428b7bb51455d1"},"source":{"id":"2605.15854","kind":"arxiv","version":1},"verdict":{"id":"ceb3fd15-e2fc-4e4b-806c-da02f3c17e8d","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T19:26:02.230336Z","strongest_claim":"Overall, we achieve a 40% relative WER decrease for Whisper and a 50% relative WER decrease for MMS, indicating that a combination of data augmentation and LLM correction is a viable strategy for the recognition of OC speech.","one_line_summary":"TTS data augmentation and LLM error correction together cut relative WER by 40-50% on ASR models for oral cancer speech.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The synthetic data produced by the augmentation techniques sufficiently captures the acoustic variability of real oral cancer speech and the LLM corrections do not systematically alter medically relevant content.","pith_extraction_headline":"Combining data augmentation and LLM error correction cuts word error rates by 40-50% for oral cancer speech recognition."},"integrity":{"clean":false,"summary":{"advisory":1,"critical":0,"by_detector":{"doi_compliance":{"total":1,"advisory":1,"critical":0,"informational":0}},"informational":0},"endpoint":"/pith/2605.15854/integrity.json","findings":[{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1044/2024JSLHR-24-00045) was visible in the surrounding text but could not be confirmed against doi.org as printed.","detector":"doi_compliance","severity":"advisory","ref_index":61,"audited_at":"2026-05-19T19:40:57.990427Z","detected_doi":"10.1044/2024JSLHR-24-00045","finding_type":"recoverable_identifier","verdict_class":"incontrovertible","detected_arxiv_id":null}],"available":true,"detectors_run":[{"name":"doi_compliance","ran_at":"2026-05-19T19:40:57.990427Z","status":"completed","version":"1.0.0","findings_count":1},{"name":"doi_title_agreement","ran_at":"2026-05-19T19:31:19.084657Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:48.705647Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T17:01:55.825308Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"869dea543c259e50062e86d21eb9b61f6989917b39c9a5a3e75fd1290202ab7d"},"references":{"count":62,"sample":[{"doi":"","year":2012,"title":"The global incidence of lip, oral cavity, and pharyngeal cancers by subsite in 2012,","work_id":"ae279b28-c2c8-49aa-b758-e5672891b000","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1002/ijc.33588","year":2020,"title":"Cancer statistics for the year 2020: An overview,","work_id":"d415b64b-0e5c-49cc-86d4-806ea0f1254d","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1007/978-3-030-04702-3","year":2019,"title":"Speech Deficits Associated with Oral and Oropharyngeal Carcinomas,","work_id":"d7222313-29a3-4c54-9413-73a0c00f73e8","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1002/9781119606987.ch22","year":2021,"title":"Speech Disorders Related to Head and Neck Cancer,","work_id":"96242b1e-935a-499f-9a5c-f97821d88b23","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1111/1460-6984.13148","year":2025,"title":"Articulatory–kinematic changes in speech following surgical treatment for oral or oropharyngeal cancer: A systematic review,","work_id":"59fd78ec-95da-45d6-90c0-3127d9afef78","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":62,"snapshot_sha256":"71fd6af53931fbdcfcd4d6f5e1fd817bdd899b91b3db06032969942c5265abbc","internal_anchors":0},"formal_canon":{"evidence_count":1,"snapshot_sha256":"eb4db4cca452aad6f9767da5791a15c852b8f374ed0dbe004eae4bb7bd58b6a9"},"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":"2605.15854","created_at":"2026-05-20T00:01:22.053982+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.15854v1","created_at":"2026-05-20T00:01:22.053982+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15854","created_at":"2026-05-20T00:01:22.053982+00:00"},{"alias_kind":"pith_short_12","alias_value":"OV4SSAHPIPZC","created_at":"2026-05-20T00:01:22.053982+00:00"},{"alias_kind":"pith_short_16","alias_value":"OV4SSAHPIPZCVVMN","created_at":"2026-05-20T00:01:22.053982+00:00"},{"alias_kind":"pith_short_8","alias_value":"OV4SSAHP","created_at":"2026-05-20T00:01:22.053982+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":1,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4","json":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4.json","graph_json":"https://pith.science/api/pith-number/OV4SSAHPIPZCVVMNDPQR7ZCQB4/graph.json","events_json":"https://pith.science/api/pith-number/OV4SSAHPIPZCVVMNDPQR7ZCQB4/events.json","paper":"https://pith.science/paper/OV4SSAHP"},"agent_actions":{"view_html":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4","download_json":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4.json","view_paper":"https://pith.science/paper/OV4SSAHP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.15854&json=true","fetch_graph":"https://pith.science/api/pith-number/OV4SSAHPIPZCVVMNDPQR7ZCQB4/graph.json","fetch_events":"https://pith.science/api/pith-number/OV4SSAHPIPZCVVMNDPQR7ZCQB4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4/action/storage_attestation","attest_author":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4/action/author_attestation","sign_citation":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4/action/citation_signature","submit_replication":"https://pith.science/pith/OV4SSAHPIPZCVVMNDPQR7ZCQB4/action/replication_record"}},"created_at":"2026-05-20T00:01:22.053982+00:00","updated_at":"2026-05-20T00:01:22.053982+00:00"}