{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:HDH6J3YPISA47Q4ONFEYH3TULJ","short_pith_number":"pith:HDH6J3YP","schema_version":"1.0","canonical_sha256":"38cfe4ef0f4481cfc38e694983ee745a7c4209c41360ce28be16ed6a23d0f93b","source":{"kind":"arxiv","id":"2203.02317","version":1},"attestation_state":"computed","paper":{"title":"Adaptive Discounting of Implicit Language Models in RNN-Transducers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Ashish Mittal, Preethi Jyothi, Samarth Bharadwaj, Shreya Khare, Sunita Sarawagi, Vinit Unni","submitted_at":"2022-02-21T08:44:56Z","abstract_excerpt":"RNN-Transducer (RNN-T) models have become synonymous with streaming end-to-end ASR systems. While they perform competitively on a number of evaluation categories, rare words pose a serious challenge to RNN-T models. One main reason for the degradation in performance on rare words is that the language model (LM) internal to RNN-Ts can become overconfident and lead to hallucinated predictions that are acoustically inconsistent with the underlying speech. To address this issue, we propose a lightweight adaptive LM discounting technique AdaptLMD, that can be used with any RNN-T architecture withou"},"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":"2203.02317","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-02-21T08:44:56Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f25e25a31d048037747221e35aa3679f0669e791a70893744efcb9b11ee894d1","abstract_canon_sha256":"c5810de4c3329cebdf9a457f8adfe33051a77c595cea70d01c13646cfa738664"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:02:03.621174Z","signature_b64":"2HxPPvwPKNNA4BaSka4X0PSMcJOGcg5omBviAKlwxknq3Nb+Oa+gQpDVRIWY+fhwEmodXf5IQumSKEK8kg9SCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"38cfe4ef0f4481cfc38e694983ee745a7c4209c41360ce28be16ed6a23d0f93b","last_reissued_at":"2026-07-05T04:02:03.620783Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:02:03.620783Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adaptive Discounting of Implicit Language Models in RNN-Transducers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Ashish Mittal, Preethi Jyothi, Samarth Bharadwaj, Shreya Khare, Sunita Sarawagi, Vinit Unni","submitted_at":"2022-02-21T08:44:56Z","abstract_excerpt":"RNN-Transducer (RNN-T) models have become synonymous with streaming end-to-end ASR systems. While they perform competitively on a number of evaluation categories, rare words pose a serious challenge to RNN-T models. One main reason for the degradation in performance on rare words is that the language model (LM) internal to RNN-Ts can become overconfident and lead to hallucinated predictions that are acoustically inconsistent with the underlying speech. To address this issue, we propose a lightweight adaptive LM discounting technique AdaptLMD, that can be used with any RNN-T architecture withou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.02317","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/2203.02317/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":"2203.02317","created_at":"2026-07-05T04:02:03.620850+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.02317v1","created_at":"2026-07-05T04:02:03.620850+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.02317","created_at":"2026-07-05T04:02:03.620850+00:00"},{"alias_kind":"pith_short_12","alias_value":"HDH6J3YPISA4","created_at":"2026-07-05T04:02:03.620850+00:00"},{"alias_kind":"pith_short_16","alias_value":"HDH6J3YPISA47Q4O","created_at":"2026-07-05T04:02:03.620850+00:00"},{"alias_kind":"pith_short_8","alias_value":"HDH6J3YP","created_at":"2026-07-05T04:02:03.620850+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ","json":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ.json","graph_json":"https://pith.science/api/pith-number/HDH6J3YPISA47Q4ONFEYH3TULJ/graph.json","events_json":"https://pith.science/api/pith-number/HDH6J3YPISA47Q4ONFEYH3TULJ/events.json","paper":"https://pith.science/paper/HDH6J3YP"},"agent_actions":{"view_html":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ","download_json":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ.json","view_paper":"https://pith.science/paper/HDH6J3YP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.02317&json=true","fetch_graph":"https://pith.science/api/pith-number/HDH6J3YPISA47Q4ONFEYH3TULJ/graph.json","fetch_events":"https://pith.science/api/pith-number/HDH6J3YPISA47Q4ONFEYH3TULJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ/action/storage_attestation","attest_author":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ/action/author_attestation","sign_citation":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ/action/citation_signature","submit_replication":"https://pith.science/pith/HDH6J3YPISA47Q4ONFEYH3TULJ/action/replication_record"}},"created_at":"2026-07-05T04:02:03.620850+00:00","updated_at":"2026-07-05T04:02:03.620850+00:00"}