{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:VXWM3Q2G73MQG3IUNFEUQBDTGD","short_pith_number":"pith:VXWM3Q2G","schema_version":"1.0","canonical_sha256":"adeccdc346fed9036d14694948047330ea6e600fb638174dd7a930dc1bf458e0","source":{"kind":"arxiv","id":"2104.02194","version":2},"attestation_state":"computed","paper":{"title":"Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusion","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","eess.AS"],"primary_cat":"cs.CL","authors_text":"Christian Fuegen, Duc Le, Gil Keren, Jay Mahadeokar, Julian Chan, Mahaveer Jain, Michael L. Seltzer, Ozlem Kalinli, Suyoun Kim, Yangyang Shi, Yatharth Saraf, Yuan Shangguan","submitted_at":"2021-04-05T23:59:43Z","abstract_excerpt":"How to leverage dynamic contextual information in end-to-end speech recognition has remained an active research area. Previous solutions to this problem were either designed for specialized use cases that did not generalize well to open-domain scenarios, did not scale to large biasing lists, or underperformed on rare long-tail words. We address these limitations by proposing a novel solution that combines shallow fusion, trie-based deep biasing, and neural network language model contextualization. These techniques result in significant 19.5% relative Word Error Rate improvement over existing c"},"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":"2104.02194","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-04-05T23:59:43Z","cross_cats_sorted":["cs.LG","eess.AS"],"title_canon_sha256":"2a2305dac1b628882e2f048396f6c293712e6ee26793ea2ddd09d790ddb8b6c5","abstract_canon_sha256":"8d075a292ed1162e66a1530e1e99a8edaa41f35e88bb6b4208d897d084fc6e0a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:48:36.898491Z","signature_b64":"EEp69k4cOAsn57P5/Z8oMlwGH39JCjS2VyLNkirIgafZdEQDBbLBmVyyHkfcqOVmfOGdBEPHimyUAuUr5Rg6CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"adeccdc346fed9036d14694948047330ea6e600fb638174dd7a930dc1bf458e0","last_reissued_at":"2026-07-05T02:48:36.898013Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:48:36.898013Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusion","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","eess.AS"],"primary_cat":"cs.CL","authors_text":"Christian Fuegen, Duc Le, Gil Keren, Jay Mahadeokar, Julian Chan, Mahaveer Jain, Michael L. Seltzer, Ozlem Kalinli, Suyoun Kim, Yangyang Shi, Yatharth Saraf, Yuan Shangguan","submitted_at":"2021-04-05T23:59:43Z","abstract_excerpt":"How to leverage dynamic contextual information in end-to-end speech recognition has remained an active research area. Previous solutions to this problem were either designed for specialized use cases that did not generalize well to open-domain scenarios, did not scale to large biasing lists, or underperformed on rare long-tail words. We address these limitations by proposing a novel solution that combines shallow fusion, trie-based deep biasing, and neural network language model contextualization. These techniques result in significant 19.5% relative Word Error Rate improvement over existing c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2104.02194","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/2104.02194/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":"2104.02194","created_at":"2026-07-05T02:48:36.898070+00:00"},{"alias_kind":"arxiv_version","alias_value":"2104.02194v2","created_at":"2026-07-05T02:48:36.898070+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2104.02194","created_at":"2026-07-05T02:48:36.898070+00:00"},{"alias_kind":"pith_short_12","alias_value":"VXWM3Q2G73MQ","created_at":"2026-07-05T02:48:36.898070+00:00"},{"alias_kind":"pith_short_16","alias_value":"VXWM3Q2G73MQG3IU","created_at":"2026-07-05T02:48:36.898070+00:00"},{"alias_kind":"pith_short_8","alias_value":"VXWM3Q2G","created_at":"2026-07-05T02:48:36.898070+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2508.07285","citing_title":"Non-Intrusive Automatic Speech Recognition Refinement: A Survey","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD","json":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD.json","graph_json":"https://pith.science/api/pith-number/VXWM3Q2G73MQG3IUNFEUQBDTGD/graph.json","events_json":"https://pith.science/api/pith-number/VXWM3Q2G73MQG3IUNFEUQBDTGD/events.json","paper":"https://pith.science/paper/VXWM3Q2G"},"agent_actions":{"view_html":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD","download_json":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD.json","view_paper":"https://pith.science/paper/VXWM3Q2G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2104.02194&json=true","fetch_graph":"https://pith.science/api/pith-number/VXWM3Q2G73MQG3IUNFEUQBDTGD/graph.json","fetch_events":"https://pith.science/api/pith-number/VXWM3Q2G73MQG3IUNFEUQBDTGD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD/action/storage_attestation","attest_author":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD/action/author_attestation","sign_citation":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD/action/citation_signature","submit_replication":"https://pith.science/pith/VXWM3Q2G73MQG3IUNFEUQBDTGD/action/replication_record"}},"created_at":"2026-07-05T02:48:36.898070+00:00","updated_at":"2026-07-05T02:48:36.898070+00:00"}