{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BZIX3N5GLCI2M6IKXE3HZF2NEN","short_pith_number":"pith:BZIX3N5G","schema_version":"1.0","canonical_sha256":"0e517db7a65891a6790ab9367c974d2348aa8ffdacb1b997fbc2e3365f4debfb","source":{"kind":"arxiv","id":"2507.21642","version":2},"attestation_state":"computed","paper":{"title":"Whilter: A Whisper-based Data Filter for \"In-the-Wild\" Speech Corpora Using Utterance-level Multi-Task Classification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"Dmitry Sityaev, George Close, Jamie Stacey, Kit Bower-Morris, Kris Y. Hong, William Ravenscroft","submitted_at":"2025-07-29T09:58:45Z","abstract_excerpt":"Large-scale in-the-wild speech datasets have become more prevalent in recent years due to increased interest in models that can learn useful features from unlabelled data for tasks such as speech recognition or synthesis. These datasets often contain undesirable features, such as multiple speakers, non-target languages, and music, which may impact model learning. The Whilter model is proposed as a multitask solution to identify these undesirable samples. Whilter uses a Whisper encoder with an attention-based classifier to solve five diverse classification problems at once. In addition, an anno"},"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":"2507.21642","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2025-07-29T09:58:45Z","cross_cats_sorted":["cs.LG","eess.AS"],"title_canon_sha256":"186c27a41770db4312f8f9028f106fc27c097fbf6b4a4468a7b6a56bdf92fe17","abstract_canon_sha256":"72d1ebd124eb6345495c6179509b9b80901d08f8e31d707bdec5b2a206548b57"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:58:33.311757Z","signature_b64":"bp4quu0IKC9PZr87YgIpXXNL5cn6JAjtisvIcVyKdRPgELU79L/6tBxKyIbNPHS6BwoeArEjZ+GE3I5oKjtpAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0e517db7a65891a6790ab9367c974d2348aa8ffdacb1b997fbc2e3365f4debfb","last_reissued_at":"2026-07-05T11:58:33.311206Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:58:33.311206Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Whilter: A Whisper-based Data Filter for \"In-the-Wild\" Speech Corpora Using Utterance-level Multi-Task Classification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"Dmitry Sityaev, George Close, Jamie Stacey, Kit Bower-Morris, Kris Y. Hong, William Ravenscroft","submitted_at":"2025-07-29T09:58:45Z","abstract_excerpt":"Large-scale in-the-wild speech datasets have become more prevalent in recent years due to increased interest in models that can learn useful features from unlabelled data for tasks such as speech recognition or synthesis. These datasets often contain undesirable features, such as multiple speakers, non-target languages, and music, which may impact model learning. The Whilter model is proposed as a multitask solution to identify these undesirable samples. Whilter uses a Whisper encoder with an attention-based classifier to solve five diverse classification problems at once. In addition, an anno"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.21642","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/2507.21642/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":"2507.21642","created_at":"2026-07-05T11:58:33.311288+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.21642v2","created_at":"2026-07-05T11:58:33.311288+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.21642","created_at":"2026-07-05T11:58:33.311288+00:00"},{"alias_kind":"pith_short_12","alias_value":"BZIX3N5GLCI2","created_at":"2026-07-05T11:58:33.311288+00:00"},{"alias_kind":"pith_short_16","alias_value":"BZIX3N5GLCI2M6IK","created_at":"2026-07-05T11:58:33.311288+00:00"},{"alias_kind":"pith_short_8","alias_value":"BZIX3N5G","created_at":"2026-07-05T11:58:33.311288+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.21642","citing_title":"Whilter: A Whisper-based Data Filter for \"In-the-Wild\" Speech Corpora Using Utterance-level Multi-Task Classification","ref_index":2,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN","json":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN.json","graph_json":"https://pith.science/api/pith-number/BZIX3N5GLCI2M6IKXE3HZF2NEN/graph.json","events_json":"https://pith.science/api/pith-number/BZIX3N5GLCI2M6IKXE3HZF2NEN/events.json","paper":"https://pith.science/paper/BZIX3N5G"},"agent_actions":{"view_html":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN","download_json":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN.json","view_paper":"https://pith.science/paper/BZIX3N5G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.21642&json=true","fetch_graph":"https://pith.science/api/pith-number/BZIX3N5GLCI2M6IKXE3HZF2NEN/graph.json","fetch_events":"https://pith.science/api/pith-number/BZIX3N5GLCI2M6IKXE3HZF2NEN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN/action/storage_attestation","attest_author":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN/action/author_attestation","sign_citation":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN/action/citation_signature","submit_replication":"https://pith.science/pith/BZIX3N5GLCI2M6IKXE3HZF2NEN/action/replication_record"}},"created_at":"2026-07-05T11:58:33.311288+00:00","updated_at":"2026-07-05T11:58:33.311288+00:00"}