{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:EV7NYA3RHHGUGFQSWNB2WDKBDC","short_pith_number":"pith:EV7NYA3R","schema_version":"1.0","canonical_sha256":"257edc037139cd431612b343ab0d4118b458cd468218aec92ff1ae3f727f1fb1","source":{"kind":"arxiv","id":"2501.16131","version":1},"attestation_state":"computed","paper":{"title":"Optimized Self-supervised Training with BEST-RQ for Speech Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.SD","authors_text":"Dominik Wagner, Ilja Baumann, Korbinian Riedhammer, Tobias Bocklet","submitted_at":"2025-01-27T15:20:50Z","abstract_excerpt":"Self-supervised learning has been successfully used for various speech related tasks, including automatic speech recognition. BERT-based Speech pre-Training with Random-projection Quantizer (BEST-RQ) has achieved state-of-the-art results in speech recognition. In this work, we further optimize the BEST-RQ approach using Kullback-Leibler divergence as an additional regularizing loss and multi-codebook extension per cluster derived from low-level feature clustering. Preliminary experiments on train-100 split of LibriSpeech result in a relative improvement of 11.2% on test-clean by using multiple"},"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":"2501.16131","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SD","submitted_at":"2025-01-27T15:20:50Z","cross_cats_sorted":[],"title_canon_sha256":"3daa5c6bd5c7d9835d49a5a50f98e835d2e6864f1c497a1716c97c37db68c5b0","abstract_canon_sha256":"034f89e17eae44064f794d600fa6454d09d4d83982c620028dd43e19d48c12d6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:05:55.804399Z","signature_b64":"lvn81qs4ecFdj31Yett1N+zM5ixkNSYQRYf5s6aIy9rAskmRO9gproqt9DRni2JlD3pmqUkvNkFVM5td2dLwAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"257edc037139cd431612b343ab0d4118b458cd468218aec92ff1ae3f727f1fb1","last_reissued_at":"2026-07-05T10:05:55.803870Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:05:55.803870Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimized Self-supervised Training with BEST-RQ for Speech Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.SD","authors_text":"Dominik Wagner, Ilja Baumann, Korbinian Riedhammer, Tobias Bocklet","submitted_at":"2025-01-27T15:20:50Z","abstract_excerpt":"Self-supervised learning has been successfully used for various speech related tasks, including automatic speech recognition. BERT-based Speech pre-Training with Random-projection Quantizer (BEST-RQ) has achieved state-of-the-art results in speech recognition. In this work, we further optimize the BEST-RQ approach using Kullback-Leibler divergence as an additional regularizing loss and multi-codebook extension per cluster derived from low-level feature clustering. Preliminary experiments on train-100 split of LibriSpeech result in a relative improvement of 11.2% on test-clean by using multiple"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.16131","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/2501.16131/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":"2501.16131","created_at":"2026-07-05T10:05:55.803933+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.16131v1","created_at":"2026-07-05T10:05:55.803933+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.16131","created_at":"2026-07-05T10:05:55.803933+00:00"},{"alias_kind":"pith_short_12","alias_value":"EV7NYA3RHHGU","created_at":"2026-07-05T10:05:55.803933+00:00"},{"alias_kind":"pith_short_16","alias_value":"EV7NYA3RHHGUGFQS","created_at":"2026-07-05T10:05:55.803933+00:00"},{"alias_kind":"pith_short_8","alias_value":"EV7NYA3R","created_at":"2026-07-05T10:05:55.803933+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC","json":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC.json","graph_json":"https://pith.science/api/pith-number/EV7NYA3RHHGUGFQSWNB2WDKBDC/graph.json","events_json":"https://pith.science/api/pith-number/EV7NYA3RHHGUGFQSWNB2WDKBDC/events.json","paper":"https://pith.science/paper/EV7NYA3R"},"agent_actions":{"view_html":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC","download_json":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC.json","view_paper":"https://pith.science/paper/EV7NYA3R","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.16131&json=true","fetch_graph":"https://pith.science/api/pith-number/EV7NYA3RHHGUGFQSWNB2WDKBDC/graph.json","fetch_events":"https://pith.science/api/pith-number/EV7NYA3RHHGUGFQSWNB2WDKBDC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC/action/storage_attestation","attest_author":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC/action/author_attestation","sign_citation":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC/action/citation_signature","submit_replication":"https://pith.science/pith/EV7NYA3RHHGUGFQSWNB2WDKBDC/action/replication_record"}},"created_at":"2026-07-05T10:05:55.803933+00:00","updated_at":"2026-07-05T10:05:55.803933+00:00"}