{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:CMLEP6DZY53XY3QCE74GRT4EE4","short_pith_number":"pith:CMLEP6DZ","schema_version":"1.0","canonical_sha256":"131647f879c7777c6e0227f868cf84272b9cacb92bdd6a493955a22752137293","source":{"kind":"arxiv","id":"2309.08876","version":2},"attestation_state":"computed","paper":{"title":"Decoder-only Architecture for Speech Recognition with CTC Prompts and Text Data Augmentation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD"],"primary_cat":"eess.AS","authors_text":"Emiru Tsunoo, Hayato Futami, Shinji Watanabe, Siddhant Arora, Yosuke Kashiwagi","submitted_at":"2023-09-16T04:57:37Z","abstract_excerpt":"Collecting audio-text pairs is expensive; however, it is much easier to access text-only data. Unless using shallow fusion, end-to-end automatic speech recognition (ASR) models require architecture modifications or additional training schemes to use text-only data. Inspired by recent advances in decoder-only language models (LMs), such as GPT-3 and PaLM adopted for speech-processing tasks, we propose using a decoder-only architecture for ASR with simple text augmentation. To provide audio information, encoder features compressed by CTC prediction are used as prompts for the decoder, which can "},"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":"2309.08876","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.AS","submitted_at":"2023-09-16T04:57:37Z","cross_cats_sorted":["cs.SD"],"title_canon_sha256":"996526acab70e8798028f0335d8b6418a80906240cecf5632d4bbf3337d22be0","abstract_canon_sha256":"2429be8c7081b91724d0af06ebabcddf4767907ac23b563d772b01cef51d095c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:31:31.950083Z","signature_b64":"vk/a44iO4afC/jO4oVVEX3ODTDQI2vuIk+JUbRc9OOeylPDx4VUez9Htkb8+F9yvcVL15xKZOQ6tVVxAlQe7AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"131647f879c7777c6e0227f868cf84272b9cacb92bdd6a493955a22752137293","last_reissued_at":"2026-07-05T07:31:31.949542Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:31:31.949542Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Decoder-only Architecture for Speech Recognition with CTC Prompts and Text Data Augmentation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD"],"primary_cat":"eess.AS","authors_text":"Emiru Tsunoo, Hayato Futami, Shinji Watanabe, Siddhant Arora, Yosuke Kashiwagi","submitted_at":"2023-09-16T04:57:37Z","abstract_excerpt":"Collecting audio-text pairs is expensive; however, it is much easier to access text-only data. Unless using shallow fusion, end-to-end automatic speech recognition (ASR) models require architecture modifications or additional training schemes to use text-only data. Inspired by recent advances in decoder-only language models (LMs), such as GPT-3 and PaLM adopted for speech-processing tasks, we propose using a decoder-only architecture for ASR with simple text augmentation. To provide audio information, encoder features compressed by CTC prediction are used as prompts for the decoder, which can "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.08876","kind":"arxiv","version":2},"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/2309.08876/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":"2309.08876","created_at":"2026-07-05T07:31:31.949598+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.08876v2","created_at":"2026-07-05T07:31:31.949598+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.08876","created_at":"2026-07-05T07:31:31.949598+00:00"},{"alias_kind":"pith_short_12","alias_value":"CMLEP6DZY53X","created_at":"2026-07-05T07:31:31.949598+00:00"},{"alias_kind":"pith_short_16","alias_value":"CMLEP6DZY53XY3QC","created_at":"2026-07-05T07:31:31.949598+00:00"},{"alias_kind":"pith_short_8","alias_value":"CMLEP6DZ","created_at":"2026-07-05T07:31:31.949598+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.12199","citing_title":"Which Speech Representation Better Matches Text-Native Reasoning? A Study of Speech-Text Alignment on Frame Rate and Representation","ref_index":53,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4","json":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4.json","graph_json":"https://pith.science/api/pith-number/CMLEP6DZY53XY3QCE74GRT4EE4/graph.json","events_json":"https://pith.science/api/pith-number/CMLEP6DZY53XY3QCE74GRT4EE4/events.json","paper":"https://pith.science/paper/CMLEP6DZ"},"agent_actions":{"view_html":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4","download_json":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4.json","view_paper":"https://pith.science/paper/CMLEP6DZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.08876&json=true","fetch_graph":"https://pith.science/api/pith-number/CMLEP6DZY53XY3QCE74GRT4EE4/graph.json","fetch_events":"https://pith.science/api/pith-number/CMLEP6DZY53XY3QCE74GRT4EE4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4/action/storage_attestation","attest_author":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4/action/author_attestation","sign_citation":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4/action/citation_signature","submit_replication":"https://pith.science/pith/CMLEP6DZY53XY3QCE74GRT4EE4/action/replication_record"}},"created_at":"2026-07-05T07:31:31.949598+00:00","updated_at":"2026-07-05T07:31:31.949598+00:00"}