{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RZFDABDCXLCHAG2H7HUZQTHEU7","short_pith_number":"pith:RZFDABDC","schema_version":"1.0","canonical_sha256":"8e4a300462bac4701b47f9e9984ce4a7e073d85e39fd7dd10f17fc0621505d05","source":{"kind":"arxiv","id":"2406.19674","version":1},"attestation_state":"computed","paper":{"title":"Less is More: Accurate Speech Recognition & Translation without Web-Scale Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Boris Ginsburg, Elena Rastorgueva, He Huang, Jagadeesh Balam, Krishna C. Puvvada, Kunal Dhawan, Nithin Rao Koluguri, Oleksii Hrinchuk, Piotr \\.Zelasko, Somshubra Majumdar, Vitaly Lavrukhin, Zhehuai Chen","submitted_at":"2024-06-28T06:22:23Z","abstract_excerpt":"Recent advances in speech recognition and translation rely on hundreds of thousands of hours of Internet speech data. We argue that state-of-the art accuracy can be reached without relying on web-scale data. Canary - multilingual ASR and speech translation model, outperforms current state-of-the-art models - Whisper, OWSM, and Seamless-M4T on English, French, Spanish, and German languages, while being trained on an order of magnitude less data than these models. Three key factors enables such data-efficient model: (1) a FastConformer-based attention encoder-decoder architecture (2) training on"},"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":"2406.19674","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-28T06:22:23Z","cross_cats_sorted":["cs.LG","cs.SD","eess.AS"],"title_canon_sha256":"8718d05b0f83b3b8b4da13ef4c90562cf2835c5fd0e8c6946a18dd305dedb404","abstract_canon_sha256":"986506a31907518236924ec1d0a652c2feb05bc96f56fe2ec3d1c040ac3973ed"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:37:43.462355Z","signature_b64":"wpEM36zWmYFGMYoBG/7BTOIFNGcTX/jEGjZMWPkNRdweliUHXrN41ebmUX6ErtX4F/j/vewzLRDvzphdklMuDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8e4a300462bac4701b47f9e9984ce4a7e073d85e39fd7dd10f17fc0621505d05","last_reissued_at":"2026-07-05T08:37:43.461813Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:37:43.461813Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Less is More: Accurate Speech Recognition & Translation without Web-Scale Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Boris Ginsburg, Elena Rastorgueva, He Huang, Jagadeesh Balam, Krishna C. Puvvada, Kunal Dhawan, Nithin Rao Koluguri, Oleksii Hrinchuk, Piotr \\.Zelasko, Somshubra Majumdar, Vitaly Lavrukhin, Zhehuai Chen","submitted_at":"2024-06-28T06:22:23Z","abstract_excerpt":"Recent advances in speech recognition and translation rely on hundreds of thousands of hours of Internet speech data. We argue that state-of-the art accuracy can be reached without relying on web-scale data. Canary - multilingual ASR and speech translation model, outperforms current state-of-the-art models - Whisper, OWSM, and Seamless-M4T on English, French, Spanish, and German languages, while being trained on an order of magnitude less data than these models. Three key factors enables such data-efficient model: (1) a FastConformer-based attention encoder-decoder architecture (2) training on"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.19674","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/2406.19674/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":"2406.19674","created_at":"2026-07-05T08:37:43.461876+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.19674v1","created_at":"2026-07-05T08:37:43.461876+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.19674","created_at":"2026-07-05T08:37:43.461876+00:00"},{"alias_kind":"pith_short_12","alias_value":"RZFDABDCXLCH","created_at":"2026-07-05T08:37:43.461876+00:00"},{"alias_kind":"pith_short_16","alias_value":"RZFDABDCXLCHAG2H","created_at":"2026-07-05T08:37:43.461876+00:00"},{"alias_kind":"pith_short_8","alias_value":"RZFDABDC","created_at":"2026-07-05T08:37:43.461876+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28732","citing_title":"CTC-Seeded Token Edit Refinement for Non-Autoregressive Speech Recognition","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23619","citing_title":"Frame-Aligned Fusion of Canary and WavLM for Non-Intrusive Intelligibility Prediction of Hearing-Aid-Processed Speech","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10736","citing_title":"BlasBench: An Open Benchmark for Irish Speech Recognition","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7","json":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7.json","graph_json":"https://pith.science/api/pith-number/RZFDABDCXLCHAG2H7HUZQTHEU7/graph.json","events_json":"https://pith.science/api/pith-number/RZFDABDCXLCHAG2H7HUZQTHEU7/events.json","paper":"https://pith.science/paper/RZFDABDC"},"agent_actions":{"view_html":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7","download_json":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7.json","view_paper":"https://pith.science/paper/RZFDABDC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.19674&json=true","fetch_graph":"https://pith.science/api/pith-number/RZFDABDCXLCHAG2H7HUZQTHEU7/graph.json","fetch_events":"https://pith.science/api/pith-number/RZFDABDCXLCHAG2H7HUZQTHEU7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7/action/storage_attestation","attest_author":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7/action/author_attestation","sign_citation":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7/action/citation_signature","submit_replication":"https://pith.science/pith/RZFDABDCXLCHAG2H7HUZQTHEU7/action/replication_record"}},"created_at":"2026-07-05T08:37:43.461876+00:00","updated_at":"2026-07-05T08:37:43.461876+00:00"}