{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:PSBUIGWSFLAFIST36U5WVABAPC","short_pith_number":"pith:PSBUIGWS","schema_version":"1.0","canonical_sha256":"7c83441ad22ac0544a7bf53b6a802078982fdabd27d1ddbeade1d670913d4445","source":{"kind":"arxiv","id":"2505.15646","version":1},"attestation_state":"computed","paper":{"title":"Word Level Timestamp Generation for Automatic Speech Recognition and Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Boris Ginsburg, Elena Rastorgueva, Hainan Xu, He Huang, Jagadeesh Balam, Ke Hu, Krishna Puvvada, Kunal Dhawan, Shuoyang Ding, Zhehuai Chen","submitted_at":"2025-05-21T15:24:29Z","abstract_excerpt":"We introduce a data-driven approach for enabling word-level timestamp prediction in the Canary model. Accurate timestamp information is crucial for a variety of downstream tasks such as speech content retrieval and timed subtitles. While traditional hybrid systems and end-to-end (E2E) models may employ external modules for timestamp prediction, our approach eliminates the need for separate alignment mechanisms. By leveraging the NeMo Forced Aligner (NFA) as a teacher model, we generate word-level timestamps and train the Canary model to predict timestamps directly. We introduce a new <|timesta"},"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":"2505.15646","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-05-21T15:24:29Z","cross_cats_sorted":["cs.SD","eess.AS"],"title_canon_sha256":"bffe1671f470531b9cf972fad52f7ab089bd967201e7cbb6d0c7493571a8343e","abstract_canon_sha256":"625696c9bb12849afbe5e11387aaf78cab00a02a97aa9602fc791882cb2ebf11"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:06:48.969336Z","signature_b64":"WH2ZaECFAYIWkEy7h28wr4FqmLM3b/ETir5TaAXGbHOUje9CdFPwVqNXVgm9Ul9UPQp2brWTVlarNJi83kBCBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7c83441ad22ac0544a7bf53b6a802078982fdabd27d1ddbeade1d670913d4445","last_reissued_at":"2026-07-05T11:06:48.968817Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:06:48.968817Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Word Level Timestamp Generation for Automatic Speech Recognition and Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Boris Ginsburg, Elena Rastorgueva, Hainan Xu, He Huang, Jagadeesh Balam, Ke Hu, Krishna Puvvada, Kunal Dhawan, Shuoyang Ding, Zhehuai Chen","submitted_at":"2025-05-21T15:24:29Z","abstract_excerpt":"We introduce a data-driven approach for enabling word-level timestamp prediction in the Canary model. Accurate timestamp information is crucial for a variety of downstream tasks such as speech content retrieval and timed subtitles. While traditional hybrid systems and end-to-end (E2E) models may employ external modules for timestamp prediction, our approach eliminates the need for separate alignment mechanisms. By leveraging the NeMo Forced Aligner (NFA) as a teacher model, we generate word-level timestamps and train the Canary model to predict timestamps directly. We introduce a new <|timesta"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.15646","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/2505.15646/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":"2505.15646","created_at":"2026-07-05T11:06:48.968882+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.15646v1","created_at":"2026-07-05T11:06:48.968882+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.15646","created_at":"2026-07-05T11:06:48.968882+00:00"},{"alias_kind":"pith_short_12","alias_value":"PSBUIGWSFLAF","created_at":"2026-07-05T11:06:48.968882+00:00"},{"alias_kind":"pith_short_16","alias_value":"PSBUIGWSFLAFIST3","created_at":"2026-07-05T11:06:48.968882+00:00"},{"alias_kind":"pith_short_8","alias_value":"PSBUIGWS","created_at":"2026-07-05T11:06:48.968882+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.30993","citing_title":"SwanVoice: Expressive Long-Form Zero-Shot Speech Synthesis for Both Monologue and Dialogue","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22817","citing_title":"In-Sync: Adaptation of Speech Aware Large Language Models for ASR with Word Level Timestamp Predictions","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC","json":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC.json","graph_json":"https://pith.science/api/pith-number/PSBUIGWSFLAFIST36U5WVABAPC/graph.json","events_json":"https://pith.science/api/pith-number/PSBUIGWSFLAFIST36U5WVABAPC/events.json","paper":"https://pith.science/paper/PSBUIGWS"},"agent_actions":{"view_html":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC","download_json":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC.json","view_paper":"https://pith.science/paper/PSBUIGWS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.15646&json=true","fetch_graph":"https://pith.science/api/pith-number/PSBUIGWSFLAFIST36U5WVABAPC/graph.json","fetch_events":"https://pith.science/api/pith-number/PSBUIGWSFLAFIST36U5WVABAPC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC/action/storage_attestation","attest_author":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC/action/author_attestation","sign_citation":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC/action/citation_signature","submit_replication":"https://pith.science/pith/PSBUIGWSFLAFIST36U5WVABAPC/action/replication_record"}},"created_at":"2026-07-05T11:06:48.968882+00:00","updated_at":"2026-07-05T11:06:48.968882+00:00"}