{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:F3WGCP7GUGPRE5QW26DDYLHKTU","short_pith_number":"pith:F3WGCP7G","schema_version":"1.0","canonical_sha256":"2eec613fe6a19f127616d7863c2cea9d0b38998bb675799b1a3e4ccfcd9c51a2","source":{"kind":"arxiv","id":"2204.12765","version":2},"attestation_state":"computed","paper":{"title":"Why does Self-Supervised Learning for Speech Recognition Benefit Speaker Recognition?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Chengyi Wang, Furu Wei, Gang Liu, Jian Wu, Jinyu Li, Peidong Wang, Sanyuan Chen, Shujie Liu, Xiangzhan Yu, Yu Wu, Zhuo Chen","submitted_at":"2022-04-27T08:35:57Z","abstract_excerpt":"Recently, self-supervised learning (SSL) has demonstrated strong performance in speaker recognition, even if the pre-training objective is designed for speech recognition. In this paper, we study which factor leads to the success of self-supervised learning on speaker-related tasks, e.g. speaker verification (SV), through a series of carefully designed experiments. Our empirical results on the Voxceleb-1 dataset suggest that the benefit of SSL to SV task is from a combination of mask speech prediction loss, data scale, and model size, while the SSL quantizer has a minor impact. We further empl"},"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":"2204.12765","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-04-27T08:35:57Z","cross_cats_sorted":["cs.SD","eess.AS"],"title_canon_sha256":"8c96091532f7c86f95861b1a8c3c0002ac5ad4db04f1d8206dff41fca9235505","abstract_canon_sha256":"5782c9dd7d860554357bf73e6bab62a893977e3cb7b563d3a77acbab70988529"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:34:53.351875Z","signature_b64":"HTTTmUU9HgRMbgJsZ+KoVWmD4reeQEfWveL8LedsIuJCqe4dB/mj4p/+46ts5kgWh5DSyv7Ql/vmdCTyamQCCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2eec613fe6a19f127616d7863c2cea9d0b38998bb675799b1a3e4ccfcd9c51a2","last_reissued_at":"2026-07-05T04:34:53.351441Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:34:53.351441Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Why does Self-Supervised Learning for Speech Recognition Benefit Speaker Recognition?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Chengyi Wang, Furu Wei, Gang Liu, Jian Wu, Jinyu Li, Peidong Wang, Sanyuan Chen, Shujie Liu, Xiangzhan Yu, Yu Wu, Zhuo Chen","submitted_at":"2022-04-27T08:35:57Z","abstract_excerpt":"Recently, self-supervised learning (SSL) has demonstrated strong performance in speaker recognition, even if the pre-training objective is designed for speech recognition. In this paper, we study which factor leads to the success of self-supervised learning on speaker-related tasks, e.g. speaker verification (SV), through a series of carefully designed experiments. Our empirical results on the Voxceleb-1 dataset suggest that the benefit of SSL to SV task is from a combination of mask speech prediction loss, data scale, and model size, while the SSL quantizer has a minor impact. We further empl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.12765","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/2204.12765/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":"2204.12765","created_at":"2026-07-05T04:34:53.351511+00:00"},{"alias_kind":"arxiv_version","alias_value":"2204.12765v2","created_at":"2026-07-05T04:34:53.351511+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.12765","created_at":"2026-07-05T04:34:53.351511+00:00"},{"alias_kind":"pith_short_12","alias_value":"F3WGCP7GUGPR","created_at":"2026-07-05T04:34:53.351511+00:00"},{"alias_kind":"pith_short_16","alias_value":"F3WGCP7GUGPRE5QW","created_at":"2026-07-05T04:34:53.351511+00:00"},{"alias_kind":"pith_short_8","alias_value":"F3WGCP7G","created_at":"2026-07-05T04:34:53.351511+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.16306","citing_title":"Layer-wise Investigation of Large-Scale Self-Supervised Music Representation Models","ref_index":20,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU","json":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU.json","graph_json":"https://pith.science/api/pith-number/F3WGCP7GUGPRE5QW26DDYLHKTU/graph.json","events_json":"https://pith.science/api/pith-number/F3WGCP7GUGPRE5QW26DDYLHKTU/events.json","paper":"https://pith.science/paper/F3WGCP7G"},"agent_actions":{"view_html":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU","download_json":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU.json","view_paper":"https://pith.science/paper/F3WGCP7G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2204.12765&json=true","fetch_graph":"https://pith.science/api/pith-number/F3WGCP7GUGPRE5QW26DDYLHKTU/graph.json","fetch_events":"https://pith.science/api/pith-number/F3WGCP7GUGPRE5QW26DDYLHKTU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU/action/storage_attestation","attest_author":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU/action/author_attestation","sign_citation":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU/action/citation_signature","submit_replication":"https://pith.science/pith/F3WGCP7GUGPRE5QW26DDYLHKTU/action/replication_record"}},"created_at":"2026-07-05T04:34:53.351511+00:00","updated_at":"2026-07-05T04:34:53.351511+00:00"}