{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:Z65QVI4GIOMSEMYDBKFOAKKB26","short_pith_number":"pith:Z65QVI4G","schema_version":"1.0","canonical_sha256":"cfbb0aa38643992233030a8ae02941d78bba49ce7b16880a8478aa9529b6d391","source":{"kind":"arxiv","id":"2012.05481","version":2},"attestation_state":"computed","paper":{"title":"Unified Streaming and Non-streaming Two-pass End-to-end Model for Speech Recognition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","eess.AS"],"primary_cat":"cs.SD","authors_text":"Binbin Zhang, Chao Yang, Di Wu, Fan Yu, Lei Xie, Liyong Guo, Xin Lei, Xiong Wang, Yaguang Hu, Zhuoyuan Yao","submitted_at":"2020-12-10T06:54:54Z","abstract_excerpt":"In this paper, we present a novel two-pass approach to unify streaming and non-streaming end-to-end (E2E) speech recognition in a single model. Our model adopts the hybrid CTC/attention architecture, in which the conformer layers in the encoder are modified. We propose a dynamic chunk-based attention strategy to allow arbitrary right context length. At inference time, the CTC decoder generates n-best hypotheses in a streaming way. The inference latency could be easily controlled by only changing the chunk size. The CTC hypotheses are then rescored by the attention decoder to get the final resu"},"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":"2012.05481","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2020-12-10T06:54:54Z","cross_cats_sorted":["cs.CL","eess.AS"],"title_canon_sha256":"fe1f56dfe68ed92ebb3bfee62cb651beb56bf863a093b5fbe00171ef253d3f33","abstract_canon_sha256":"20a7dfa16f68fa12b5c1dab6cf7e2aad1bd1005b198682048647210c044aabf0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:44:08.285438Z","signature_b64":"5c4yZP6jedd9EpHiMWPJUQjsRycarglVc5mBfycU4NHFOTeghTSj01KHROOS2Rz0KAmY9JDXaN0gaDpxZiDUBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cfbb0aa38643992233030a8ae02941d78bba49ce7b16880a8478aa9529b6d391","last_reissued_at":"2026-07-05T03:44:08.284996Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:44:08.284996Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unified Streaming and Non-streaming Two-pass End-to-end Model for Speech Recognition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","eess.AS"],"primary_cat":"cs.SD","authors_text":"Binbin Zhang, Chao Yang, Di Wu, Fan Yu, Lei Xie, Liyong Guo, Xin Lei, Xiong Wang, Yaguang Hu, Zhuoyuan Yao","submitted_at":"2020-12-10T06:54:54Z","abstract_excerpt":"In this paper, we present a novel two-pass approach to unify streaming and non-streaming end-to-end (E2E) speech recognition in a single model. Our model adopts the hybrid CTC/attention architecture, in which the conformer layers in the encoder are modified. We propose a dynamic chunk-based attention strategy to allow arbitrary right context length. At inference time, the CTC decoder generates n-best hypotheses in a streaming way. The inference latency could be easily controlled by only changing the chunk size. The CTC hypotheses are then rescored by the attention decoder to get the final resu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.05481","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/2012.05481/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":"2012.05481","created_at":"2026-07-05T03:44:08.285055+00:00"},{"alias_kind":"arxiv_version","alias_value":"2012.05481v2","created_at":"2026-07-05T03:44:08.285055+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.05481","created_at":"2026-07-05T03:44:08.285055+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z65QVI4GIOMS","created_at":"2026-07-05T03:44:08.285055+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z65QVI4GIOMSEMYD","created_at":"2026-07-05T03:44:08.285055+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z65QVI4G","created_at":"2026-07-05T03:44:08.285055+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.18105","citing_title":"NIM4-ASR: Towards Efficient, Robust, and Customizable Real-Time LLM-Based ASR","ref_index":31,"is_internal_anchor":true},{"citing_arxiv_id":"2606.21268","citing_title":"Online Predictive Coding for Dual-Mode Self-Supervised Speech Model","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08486","citing_title":"TRADE: Transducer-Augmented Decoder for Speech LLM","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2601.20896","citing_title":"A Study of Data Selection Strategies for Pre-training Self-Supervised Speech Models","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18105","citing_title":"NIM4-ASR: Towards Efficient, Robust, and Customizable Real-Time LLM-Based ASR","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26","json":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26.json","graph_json":"https://pith.science/api/pith-number/Z65QVI4GIOMSEMYDBKFOAKKB26/graph.json","events_json":"https://pith.science/api/pith-number/Z65QVI4GIOMSEMYDBKFOAKKB26/events.json","paper":"https://pith.science/paper/Z65QVI4G"},"agent_actions":{"view_html":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26","download_json":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26.json","view_paper":"https://pith.science/paper/Z65QVI4G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2012.05481&json=true","fetch_graph":"https://pith.science/api/pith-number/Z65QVI4GIOMSEMYDBKFOAKKB26/graph.json","fetch_events":"https://pith.science/api/pith-number/Z65QVI4GIOMSEMYDBKFOAKKB26/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26/action/storage_attestation","attest_author":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26/action/author_attestation","sign_citation":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26/action/citation_signature","submit_replication":"https://pith.science/pith/Z65QVI4GIOMSEMYDBKFOAKKB26/action/replication_record"}},"created_at":"2026-07-05T03:44:08.285055+00:00","updated_at":"2026-07-05T03:44:08.285055+00:00"}