{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:EYKAFT7V45DB4WV3L7MMQKNCLU","short_pith_number":"pith:EYKAFT7V","schema_version":"1.0","canonical_sha256":"261402cff5e7461e5abb5fd8c829a25d2883aa52197b04a57912048463feb782","source":{"kind":"arxiv","id":"2402.02302","version":1},"attestation_state":"computed","paper":{"title":"Predicting positive transfer for improved low-resource speech recognition using acoustic pseudo-tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"eess.AS","authors_text":"Aryaman Arora, Dan Jurafsky, Georgios Paraskevopoulos, Nay San, Oliver Adams, Prabhjot Kaur, Xiluo He","submitted_at":"2024-02-03T23:54:03Z","abstract_excerpt":"While massively multilingual speech models like wav2vec 2.0 XLSR-128 can be directly fine-tuned for automatic speech recognition (ASR), downstream performance can still be relatively poor on languages that are under-represented in the pre-training data. Continued pre-training on 70-200 hours of untranscribed speech in these languages can help -- but what about languages without that much recorded data? For such cases, we show that supplementing the target language with data from a similar, higher-resource 'donor' language can help. For example, continued pre-training on only 10 hours of low-re"},"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":"2402.02302","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2024-02-03T23:54:03Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"9d4abcf8cdeb0e06d14de33d9e03c40f181f0c8d9ca3c2ade553fc0890f33121","abstract_canon_sha256":"b7aef93bfc2b03607e84904822787c571824b36358771fe93ce3d05198df233c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:41:13.182488Z","signature_b64":"VHBXymMyhcUjJkugGolEnv5Mpjda0HO7tush2zP3AVKjaV/tEcSr9IRuXP++MCVljAWTqypmdcN8/LnPKjB0CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"261402cff5e7461e5abb5fd8c829a25d2883aa52197b04a57912048463feb782","last_reissued_at":"2026-07-05T07:41:13.182051Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:41:13.182051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Predicting positive transfer for improved low-resource speech recognition using acoustic pseudo-tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"eess.AS","authors_text":"Aryaman Arora, Dan Jurafsky, Georgios Paraskevopoulos, Nay San, Oliver Adams, Prabhjot Kaur, Xiluo He","submitted_at":"2024-02-03T23:54:03Z","abstract_excerpt":"While massively multilingual speech models like wav2vec 2.0 XLSR-128 can be directly fine-tuned for automatic speech recognition (ASR), downstream performance can still be relatively poor on languages that are under-represented in the pre-training data. Continued pre-training on 70-200 hours of untranscribed speech in these languages can help -- but what about languages without that much recorded data? For such cases, we show that supplementing the target language with data from a similar, higher-resource 'donor' language can help. For example, continued pre-training on only 10 hours of low-re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.02302","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/2402.02302/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":"2402.02302","created_at":"2026-07-05T07:41:13.182113+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.02302v1","created_at":"2026-07-05T07:41:13.182113+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.02302","created_at":"2026-07-05T07:41:13.182113+00:00"},{"alias_kind":"pith_short_12","alias_value":"EYKAFT7V45DB","created_at":"2026-07-05T07:41:13.182113+00:00"},{"alias_kind":"pith_short_16","alias_value":"EYKAFT7V45DB4WV3","created_at":"2026-07-05T07:41:13.182113+00:00"},{"alias_kind":"pith_short_8","alias_value":"EYKAFT7V","created_at":"2026-07-05T07:41:13.182113+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.09206","citing_title":"SimClass: A Classroom Speech Dataset Generated via Game Engine Simulation For Automatic Speech Recognition Research","ref_index":8,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU","json":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU.json","graph_json":"https://pith.science/api/pith-number/EYKAFT7V45DB4WV3L7MMQKNCLU/graph.json","events_json":"https://pith.science/api/pith-number/EYKAFT7V45DB4WV3L7MMQKNCLU/events.json","paper":"https://pith.science/paper/EYKAFT7V"},"agent_actions":{"view_html":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU","download_json":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU.json","view_paper":"https://pith.science/paper/EYKAFT7V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.02302&json=true","fetch_graph":"https://pith.science/api/pith-number/EYKAFT7V45DB4WV3L7MMQKNCLU/graph.json","fetch_events":"https://pith.science/api/pith-number/EYKAFT7V45DB4WV3L7MMQKNCLU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU/action/storage_attestation","attest_author":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU/action/author_attestation","sign_citation":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU/action/citation_signature","submit_replication":"https://pith.science/pith/EYKAFT7V45DB4WV3L7MMQKNCLU/action/replication_record"}},"created_at":"2026-07-05T07:41:13.182113+00:00","updated_at":"2026-07-05T07:41:13.182113+00:00"}