{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:IFL7GS3QJXB3DCS7YMVCGYT5YT","short_pith_number":"pith:IFL7GS3Q","schema_version":"1.0","canonical_sha256":"4157f34b704dc3b18a5fc32a23627dc4dd42bac338add3d78c77b69f084f7d84","source":{"kind":"arxiv","id":"2112.07254","version":1},"attestation_state":"computed","paper":{"title":"Improving Hybrid CTC/Attention End-to-end Speech Recognition with Pretrained Acoustic and Language Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.SD"],"primary_cat":"eess.AS","authors_text":"Keqi Deng, Long Ma, Songjun Cao, Yike Zhang","submitted_at":"2021-12-14T09:38:31Z","abstract_excerpt":"Recently, self-supervised pretraining has achieved impressive results in end-to-end (E2E) automatic speech recognition (ASR). However, the dominant sequence-to-sequence (S2S) E2E model is still hard to fully utilize the self-supervised pre-training methods because its decoder is conditioned on acoustic representation thus cannot be pretrained separately. In this paper, we propose a pretrained Transformer (Preformer) S2S ASR architecture based on hybrid CTC/attention E2E models to fully utilize the pretrained acoustic models (AMs) and language models (LMs). In our framework, the encoder is init"},"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":"2112.07254","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.AS","submitted_at":"2021-12-14T09:38:31Z","cross_cats_sorted":["cs.CL","cs.SD"],"title_canon_sha256":"2d03165e74d5c7b13d773e6c4b575422394596d51cae44b8a452e275328fca12","abstract_canon_sha256":"3cfc4438b9f1ac9a6f01fc42ae17c899daff99ff1b49178039406c8473f8ce53"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:40:56.376637Z","signature_b64":"JL4vfoF0ZULbQZG3V/Kf7nu04LH+X9hcoi++8u0JjWA8QQP9USpxb1JMzY6Mh4hmWNmLwfGic41MKmWBVDN7AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4157f34b704dc3b18a5fc32a23627dc4dd42bac338add3d78c77b69f084f7d84","last_reissued_at":"2026-07-05T03:40:56.376150Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:40:56.376150Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Hybrid CTC/Attention End-to-end Speech Recognition with Pretrained Acoustic and Language Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.SD"],"primary_cat":"eess.AS","authors_text":"Keqi Deng, Long Ma, Songjun Cao, Yike Zhang","submitted_at":"2021-12-14T09:38:31Z","abstract_excerpt":"Recently, self-supervised pretraining has achieved impressive results in end-to-end (E2E) automatic speech recognition (ASR). However, the dominant sequence-to-sequence (S2S) E2E model is still hard to fully utilize the self-supervised pre-training methods because its decoder is conditioned on acoustic representation thus cannot be pretrained separately. In this paper, we propose a pretrained Transformer (Preformer) S2S ASR architecture based on hybrid CTC/attention E2E models to fully utilize the pretrained acoustic models (AMs) and language models (LMs). In our framework, the encoder is init"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2112.07254","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/2112.07254/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":"2112.07254","created_at":"2026-07-05T03:40:56.376207+00:00"},{"alias_kind":"arxiv_version","alias_value":"2112.07254v1","created_at":"2026-07-05T03:40:56.376207+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2112.07254","created_at":"2026-07-05T03:40:56.376207+00:00"},{"alias_kind":"pith_short_12","alias_value":"IFL7GS3QJXB3","created_at":"2026-07-05T03:40:56.376207+00:00"},{"alias_kind":"pith_short_16","alias_value":"IFL7GS3QJXB3DCS7","created_at":"2026-07-05T03:40:56.376207+00:00"},{"alias_kind":"pith_short_8","alias_value":"IFL7GS3Q","created_at":"2026-07-05T03:40:56.376207+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.08384","citing_title":"TASU2: Controllable CTC Simulation for Alignment and Low-Resource Adaptation of Speech LLMs","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT","json":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT.json","graph_json":"https://pith.science/api/pith-number/IFL7GS3QJXB3DCS7YMVCGYT5YT/graph.json","events_json":"https://pith.science/api/pith-number/IFL7GS3QJXB3DCS7YMVCGYT5YT/events.json","paper":"https://pith.science/paper/IFL7GS3Q"},"agent_actions":{"view_html":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT","download_json":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT.json","view_paper":"https://pith.science/paper/IFL7GS3Q","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2112.07254&json=true","fetch_graph":"https://pith.science/api/pith-number/IFL7GS3QJXB3DCS7YMVCGYT5YT/graph.json","fetch_events":"https://pith.science/api/pith-number/IFL7GS3QJXB3DCS7YMVCGYT5YT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT/action/storage_attestation","attest_author":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT/action/author_attestation","sign_citation":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT/action/citation_signature","submit_replication":"https://pith.science/pith/IFL7GS3QJXB3DCS7YMVCGYT5YT/action/replication_record"}},"created_at":"2026-07-05T03:40:56.376207+00:00","updated_at":"2026-07-05T03:40:56.376207+00:00"}