{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:LLT7XBLY2T3KTAWGSZAAAI6333","short_pith_number":"pith:LLT7XBLY","schema_version":"1.0","canonical_sha256":"5ae7fb8578d4f6a982c696400023dbdef2cb2e3ddea3db5336ac8c8376852a34","source":{"kind":"arxiv","id":"2506.18843","version":2},"attestation_state":"computed","paper":{"title":"USAD: Universal Speech and Audio Representation via Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","eess.AS"],"primary_cat":"cs.SD","authors_text":"Alexander H. Liu, Heng-Jui Chang, James Glass, Saurabhchand Bhati","submitted_at":"2025-06-23T17:02:00Z","abstract_excerpt":"Self-supervised learning (SSL) has revolutionized audio representations, yet models often remain domain-specific, focusing on either speech or non-speech tasks. In this work, we present Universal Speech and Audio Distillation (USAD), a unified approach to audio representation learning that integrates diverse audio types - speech, sound, and music - into a single model. USAD employs efficient layer-to-layer distillation from domain-specific SSL models to train a student on a comprehensive audio dataset. USAD offers competitive performance across various benchmarks and datasets, including frame "},"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":"2506.18843","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SD","submitted_at":"2025-06-23T17:02:00Z","cross_cats_sorted":["cs.CL","eess.AS"],"title_canon_sha256":"f89debe6f2ba1ef7029a55c855e81a1f5a88101714405c1d9bae6555b45c49ef","abstract_canon_sha256":"17a5dcd377ba79b3fda68192a3a4aef3e368cd017cb552b10afe07aa32955621"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:55:20.060794Z","signature_b64":"QhZ+D6iHfieKLrg1QSsrkLmjtZ5uKuhIFTtnoS76ypUb1cmlC83zIcLzdHyxmzJZrXaiRCwUhMHL22g4R2lOCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5ae7fb8578d4f6a982c696400023dbdef2cb2e3ddea3db5336ac8c8376852a34","last_reissued_at":"2026-07-05T11:55:20.060215Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:55:20.060215Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"USAD: Universal Speech and Audio Representation via Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","eess.AS"],"primary_cat":"cs.SD","authors_text":"Alexander H. Liu, Heng-Jui Chang, James Glass, Saurabhchand Bhati","submitted_at":"2025-06-23T17:02:00Z","abstract_excerpt":"Self-supervised learning (SSL) has revolutionized audio representations, yet models often remain domain-specific, focusing on either speech or non-speech tasks. In this work, we present Universal Speech and Audio Distillation (USAD), a unified approach to audio representation learning that integrates diverse audio types - speech, sound, and music - into a single model. USAD employs efficient layer-to-layer distillation from domain-specific SSL models to train a student on a comprehensive audio dataset. USAD offers competitive performance across various benchmarks and datasets, including frame "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.18843","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/2506.18843/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":"2506.18843","created_at":"2026-07-05T11:55:20.060282+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.18843v2","created_at":"2026-07-05T11:55:20.060282+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.18843","created_at":"2026-07-05T11:55:20.060282+00:00"},{"alias_kind":"pith_short_12","alias_value":"LLT7XBLY2T3K","created_at":"2026-07-05T11:55:20.060282+00:00"},{"alias_kind":"pith_short_16","alias_value":"LLT7XBLY2T3KTAWG","created_at":"2026-07-05T11:55:20.060282+00:00"},{"alias_kind":"pith_short_8","alias_value":"LLT7XBLY","created_at":"2026-07-05T11:55:20.060282+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.04547","citing_title":"Stage-adaptive audio diffusion modeling","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00251","citing_title":"Alethia: A Foundational Encoder for Voice Deepfakes","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333","json":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333.json","graph_json":"https://pith.science/api/pith-number/LLT7XBLY2T3KTAWGSZAAAI6333/graph.json","events_json":"https://pith.science/api/pith-number/LLT7XBLY2T3KTAWGSZAAAI6333/events.json","paper":"https://pith.science/paper/LLT7XBLY"},"agent_actions":{"view_html":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333","download_json":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333.json","view_paper":"https://pith.science/paper/LLT7XBLY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.18843&json=true","fetch_graph":"https://pith.science/api/pith-number/LLT7XBLY2T3KTAWGSZAAAI6333/graph.json","fetch_events":"https://pith.science/api/pith-number/LLT7XBLY2T3KTAWGSZAAAI6333/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333/action/storage_attestation","attest_author":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333/action/author_attestation","sign_citation":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333/action/citation_signature","submit_replication":"https://pith.science/pith/LLT7XBLY2T3KTAWGSZAAAI6333/action/replication_record"}},"created_at":"2026-07-05T11:55:20.060282+00:00","updated_at":"2026-07-05T11:55:20.060282+00:00"}