{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:Y4BLRL74USIVIR7L362OHVOEXX","short_pith_number":"pith:Y4BLRL74","schema_version":"1.0","canonical_sha256":"c702b8affca4915447ebdfb4e3d5c4bdd6c1e9f9f724091de7df24df7694475b","source":{"kind":"arxiv","id":"2408.16589","version":1},"attestation_state":"computed","paper":{"title":"CrisperWhisper: Accurate Timestamps on Verbatim Speech Transcriptions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bernhard Thallinger, Laurin Wagner, Mario Zusag","submitted_at":"2024-08-29T14:52:42Z","abstract_excerpt":"We demonstrate that carefully adjusting the tokenizer of the Whisper speech recognition model significantly improves the precision of word-level timestamps when applying dynamic time warping to the decoder's cross-attention scores. We fine-tune the model to produce more verbatim speech transcriptions and employ several techniques to increase robustness against multiple speakers and background noise. These adjustments achieve state-of-the-art performance on benchmarks for verbatim speech transcription, word segmentation, and the timed detection of filler events, and can further mitigate transcr"},"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":"2408.16589","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-29T14:52:42Z","cross_cats_sorted":[],"title_canon_sha256":"b76b2968b16f3d06bdc11b5347d6416702bdef96eddfc979e907e8cd03b16b11","abstract_canon_sha256":"92af256f99be5a43cdc5eba7961a32c10c56871fdfa95e89b3162473a2b92c62"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:00:43.475483Z","signature_b64":"3j40Ephr1bVLmkc9kbSH5+gIu6sUBe1X1uJxpj0InRh+F6vfLUfnPlNiJRlpVxcHMQaeKzRT/6ynYd1+TrFYAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c702b8affca4915447ebdfb4e3d5c4bdd6c1e9f9f724091de7df24df7694475b","last_reissued_at":"2026-07-05T09:00:43.475093Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:00:43.475093Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CrisperWhisper: Accurate Timestamps on Verbatim Speech Transcriptions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bernhard Thallinger, Laurin Wagner, Mario Zusag","submitted_at":"2024-08-29T14:52:42Z","abstract_excerpt":"We demonstrate that carefully adjusting the tokenizer of the Whisper speech recognition model significantly improves the precision of word-level timestamps when applying dynamic time warping to the decoder's cross-attention scores. We fine-tune the model to produce more verbatim speech transcriptions and employ several techniques to increase robustness against multiple speakers and background noise. These adjustments achieve state-of-the-art performance on benchmarks for verbatim speech transcription, word segmentation, and the timed detection of filler events, and can further mitigate transcr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.16589","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/2408.16589/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":"2408.16589","created_at":"2026-07-05T09:00:43.475143+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.16589v1","created_at":"2026-07-05T09:00:43.475143+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.16589","created_at":"2026-07-05T09:00:43.475143+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y4BLRL74USIV","created_at":"2026-07-05T09:00:43.475143+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y4BLRL74USIVIR7L","created_at":"2026-07-05T09:00:43.475143+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y4BLRL74","created_at":"2026-07-05T09:00:43.475143+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05364","citing_title":"REDDIT: Correcting Model-Generated Timestamp Drift in ASR without Forgetting via Replay-Based Distribution Editing","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2605.01381","citing_title":"A framework for analyzing concept representations in neural models","ref_index":168,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX","json":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX.json","graph_json":"https://pith.science/api/pith-number/Y4BLRL74USIVIR7L362OHVOEXX/graph.json","events_json":"https://pith.science/api/pith-number/Y4BLRL74USIVIR7L362OHVOEXX/events.json","paper":"https://pith.science/paper/Y4BLRL74"},"agent_actions":{"view_html":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX","download_json":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX.json","view_paper":"https://pith.science/paper/Y4BLRL74","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.16589&json=true","fetch_graph":"https://pith.science/api/pith-number/Y4BLRL74USIVIR7L362OHVOEXX/graph.json","fetch_events":"https://pith.science/api/pith-number/Y4BLRL74USIVIR7L362OHVOEXX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX/action/storage_attestation","attest_author":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX/action/author_attestation","sign_citation":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX/action/citation_signature","submit_replication":"https://pith.science/pith/Y4BLRL74USIVIR7L362OHVOEXX/action/replication_record"}},"created_at":"2026-07-05T09:00:43.475143+00:00","updated_at":"2026-07-05T09:00:43.475143+00:00"}