{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:I2T575NUEHSD56UMKN3ELNBQ2Y","short_pith_number":"pith:I2T575NU","schema_version":"1.0","canonical_sha256":"46a7dff5b421e43efa8c537645b430d6046d7599a94ba5fce49c4764ab62115f","source":{"kind":"arxiv","id":"2312.17279","version":3},"attestation_state":"computed","paper":{"title":"Stateful Conformer with Cache-based Inference for Streaming Automatic Speech Recognition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.CL","authors_text":"Ankur Kumar, Boris Ginsburg, Jagadeesh Balam, Somshubra Majumdar, Vahid Noroozi","submitted_at":"2023-12-27T21:04:26Z","abstract_excerpt":"In this paper, we propose an efficient and accurate streaming speech recognition model based on the FastConformer architecture. We adapted the FastConformer architecture for streaming applications through: (1) constraining both the look-ahead and past contexts in the encoder, and (2) introducing an activation caching mechanism to enable the non-autoregressive encoder to operate autoregressively during inference. The proposed model is thoughtfully designed in a way to eliminate the accuracy disparity between the train and inference time which is common for many streaming models. Furthermore, ou"},"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":"2312.17279","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-12-27T21:04:26Z","cross_cats_sorted":["eess.AS"],"title_canon_sha256":"eaea8fd153cd3b8b0cb9b3c5af55e4f2c0c0089bc011ef75d6db09cf6abc7265","abstract_canon_sha256":"fcd8b5eb8897f5d98dcc1ce100c9f164598e2b1326ad54baf880b8a4882bec5b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:14:52.284697Z","signature_b64":"3v9suqIce+bVy/XjqpmB6dUGgLo/yxxSdE/PDV32hIA8UuBG9gL94UUfrCbeFOdGQjiFPTQlQBLkGB5wIo+WCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"46a7dff5b421e43efa8c537645b430d6046d7599a94ba5fce49c4764ab62115f","last_reissued_at":"2026-07-05T08:14:52.284294Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:14:52.284294Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stateful Conformer with Cache-based Inference for Streaming Automatic Speech Recognition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.CL","authors_text":"Ankur Kumar, Boris Ginsburg, Jagadeesh Balam, Somshubra Majumdar, Vahid Noroozi","submitted_at":"2023-12-27T21:04:26Z","abstract_excerpt":"In this paper, we propose an efficient and accurate streaming speech recognition model based on the FastConformer architecture. We adapted the FastConformer architecture for streaming applications through: (1) constraining both the look-ahead and past contexts in the encoder, and (2) introducing an activation caching mechanism to enable the non-autoregressive encoder to operate autoregressively during inference. The proposed model is thoughtfully designed in a way to eliminate the accuracy disparity between the train and inference time which is common for many streaming models. Furthermore, ou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.17279","kind":"arxiv","version":3},"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/2312.17279/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":"2312.17279","created_at":"2026-07-05T08:14:52.284361+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.17279v3","created_at":"2026-07-05T08:14:52.284361+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.17279","created_at":"2026-07-05T08:14:52.284361+00:00"},{"alias_kind":"pith_short_12","alias_value":"I2T575NUEHSD","created_at":"2026-07-05T08:14:52.284361+00:00"},{"alias_kind":"pith_short_16","alias_value":"I2T575NUEHSD56UM","created_at":"2026-07-05T08:14:52.284361+00:00"},{"alias_kind":"pith_short_8","alias_value":"I2T575NU","created_at":"2026-07-05T08:14:52.284361+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19688","citing_title":"Latency-Configurable Streaming Speech Enhancement via Asymmetric Temporal Padding","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2602.11298","citing_title":"Voxtral Realtime","ref_index":18,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y","json":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y.json","graph_json":"https://pith.science/api/pith-number/I2T575NUEHSD56UMKN3ELNBQ2Y/graph.json","events_json":"https://pith.science/api/pith-number/I2T575NUEHSD56UMKN3ELNBQ2Y/events.json","paper":"https://pith.science/paper/I2T575NU"},"agent_actions":{"view_html":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y","download_json":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y.json","view_paper":"https://pith.science/paper/I2T575NU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.17279&json=true","fetch_graph":"https://pith.science/api/pith-number/I2T575NUEHSD56UMKN3ELNBQ2Y/graph.json","fetch_events":"https://pith.science/api/pith-number/I2T575NUEHSD56UMKN3ELNBQ2Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y/action/storage_attestation","attest_author":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y/action/author_attestation","sign_citation":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y/action/citation_signature","submit_replication":"https://pith.science/pith/I2T575NUEHSD56UMKN3ELNBQ2Y/action/replication_record"}},"created_at":"2026-07-05T08:14:52.284361+00:00","updated_at":"2026-07-05T08:14:52.284361+00:00"}