{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:ST3W25ZC7QXAFULLBISQXWKUD4","short_pith_number":"pith:ST3W25ZC","schema_version":"1.0","canonical_sha256":"94f76d7722fc2e02d16b0a250bd9541f36729ed55240ce1ebc89cdc2c99a34c3","source":{"kind":"arxiv","id":"2005.11185","version":2},"attestation_state":"computed","paper":{"title":"Low-Latency Sequence-to-Sequence Speech Recognition and Translation by Partial Hypothesis Selection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Danni Liu, Gerasimos Spanakis, Jan Niehues","submitted_at":"2020-05-22T13:42:54Z","abstract_excerpt":"Encoder-decoder models provide a generic architecture for sequence-to-sequence tasks such as speech recognition and translation. While offline systems are often evaluated on quality metrics like word error rates (WER) and BLEU, latency is also a crucial factor in many practical use-cases. We propose three latency reduction techniques for chunk-based incremental inference and evaluate their efficiency in terms of accuracy-latency trade-off. On the 300-hour How2 dataset, we reduce latency by 83% to 0.8 second by sacrificing 1% WER (6% rel.) compared to offline transcription. Although our experim"},"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":"2005.11185","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-05-22T13:42:54Z","cross_cats_sorted":["cs.SD","eess.AS"],"title_canon_sha256":"1db1349453dbc78c9a1bdcc1f1b536e26cdf13420c65ca565f51b0f67534b965","abstract_canon_sha256":"8620d35be0a78ff6cae48b40491dba1440ee9daca7aa2261adcede82a2bd2f15"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:42:36.833593Z","signature_b64":"boi1WzwieZQlfFnL8rpTDnAapPA/ehCET2e1HL6I/FeV/wxrnntoz10qQNCOaxZ+IAExllt1ldcjUkOd1sqHBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"94f76d7722fc2e02d16b0a250bd9541f36729ed55240ce1ebc89cdc2c99a34c3","last_reissued_at":"2026-07-05T01:42:36.833079Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:42:36.833079Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Low-Latency Sequence-to-Sequence Speech Recognition and Translation by Partial Hypothesis Selection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Danni Liu, Gerasimos Spanakis, Jan Niehues","submitted_at":"2020-05-22T13:42:54Z","abstract_excerpt":"Encoder-decoder models provide a generic architecture for sequence-to-sequence tasks such as speech recognition and translation. While offline systems are often evaluated on quality metrics like word error rates (WER) and BLEU, latency is also a crucial factor in many practical use-cases. We propose three latency reduction techniques for chunk-based incremental inference and evaluate their efficiency in terms of accuracy-latency trade-off. On the 300-hour How2 dataset, we reduce latency by 83% to 0.8 second by sacrificing 1% WER (6% rel.) compared to offline transcription. Although our experim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2005.11185","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/2005.11185/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":"2005.11185","created_at":"2026-07-05T01:42:36.833138+00:00"},{"alias_kind":"arxiv_version","alias_value":"2005.11185v2","created_at":"2026-07-05T01:42:36.833138+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2005.11185","created_at":"2026-07-05T01:42:36.833138+00:00"},{"alias_kind":"pith_short_12","alias_value":"ST3W25ZC7QXA","created_at":"2026-07-05T01:42:36.833138+00:00"},{"alias_kind":"pith_short_16","alias_value":"ST3W25ZC7QXAFULL","created_at":"2026-07-05T01:42:36.833138+00:00"},{"alias_kind":"pith_short_8","alias_value":"ST3W25ZC","created_at":"2026-07-05T01:42:36.833138+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"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":14,"is_internal_anchor":true},{"citing_arxiv_id":"2607.01108","citing_title":"NPUsper: Eliminating Redundant Computation for Real-Time Whisper on Mobile NPUs","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2508.12301","citing_title":"WhisperRT -- Turning Whisper into a Causal Streaming Model","ref_index":20,"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":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4","json":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4.json","graph_json":"https://pith.science/api/pith-number/ST3W25ZC7QXAFULLBISQXWKUD4/graph.json","events_json":"https://pith.science/api/pith-number/ST3W25ZC7QXAFULLBISQXWKUD4/events.json","paper":"https://pith.science/paper/ST3W25ZC"},"agent_actions":{"view_html":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4","download_json":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4.json","view_paper":"https://pith.science/paper/ST3W25ZC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2005.11185&json=true","fetch_graph":"https://pith.science/api/pith-number/ST3W25ZC7QXAFULLBISQXWKUD4/graph.json","fetch_events":"https://pith.science/api/pith-number/ST3W25ZC7QXAFULLBISQXWKUD4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4/action/storage_attestation","attest_author":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4/action/author_attestation","sign_citation":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4/action/citation_signature","submit_replication":"https://pith.science/pith/ST3W25ZC7QXAFULLBISQXWKUD4/action/replication_record"}},"created_at":"2026-07-05T01:42:36.833138+00:00","updated_at":"2026-07-05T01:42:36.833138+00:00"}