{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:AXS4FXFNIBAEOZ7KGYGAOU2XOW","short_pith_number":"pith:AXS4FXFN","schema_version":"1.0","canonical_sha256":"05e5c2dcad40404767ea360c07535775925542276018b1191fc7e6e393ebea77","source":{"kind":"arxiv","id":"2508.10949","version":1},"attestation_state":"computed","paper":{"title":"Perturbed Public Voices (P$^{2}$V): A Dataset for Robust Audio Deepfake Detection","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.SD","authors_text":"Chongyang Gao, Isabel Gortner, Marco Postiglione, Sarit Kraus, V.S. Subrahmanian","submitted_at":"2025-08-13T17:54:55Z","abstract_excerpt":"Current audio deepfake detectors cannot be trusted. While they excel on controlled benchmarks, they fail when tested in the real world. We introduce Perturbed Public Voices (P$^{2}$V), an IRB-approved dataset capturing three critical aspects of malicious deepfakes: (1) identity-consistent transcripts via LLMs, (2) environmental and adversarial noise, and (3) state-of-the-art voice cloning (2020-2025). Experiments reveal alarming vulnerabilities of 22 recent audio deepfake detectors: models trained on current datasets lose 43% performance when tested on P$^{2}$V, with performance measured as th"},"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":"2508.10949","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.SD","submitted_at":"2025-08-13T17:54:55Z","cross_cats_sorted":["eess.AS"],"title_canon_sha256":"4aa723268519e860ffbbe1a099387bb162a9f058e9735f58c75c04f1bc969a3e","abstract_canon_sha256":"47a006bee5b3fe1a4cba7a10b8cc2c28b9148420ee407221425d6a872c3f9683"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:54:16.048128Z","signature_b64":"oLcHPEj53cyZYLrTzkoNWq2EdpIQiByjFwk6WDxIzzxBEsI1e5Xu/IEuNhuckBGvj6GS2+Z2mxViy6f4XFF+Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"05e5c2dcad40404767ea360c07535775925542276018b1191fc7e6e393ebea77","last_reissued_at":"2026-07-05T11:54:16.047644Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:54:16.047644Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Perturbed Public Voices (P$^{2}$V): A Dataset for Robust Audio Deepfake Detection","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.SD","authors_text":"Chongyang Gao, Isabel Gortner, Marco Postiglione, Sarit Kraus, V.S. Subrahmanian","submitted_at":"2025-08-13T17:54:55Z","abstract_excerpt":"Current audio deepfake detectors cannot be trusted. While they excel on controlled benchmarks, they fail when tested in the real world. We introduce Perturbed Public Voices (P$^{2}$V), an IRB-approved dataset capturing three critical aspects of malicious deepfakes: (1) identity-consistent transcripts via LLMs, (2) environmental and adversarial noise, and (3) state-of-the-art voice cloning (2020-2025). Experiments reveal alarming vulnerabilities of 22 recent audio deepfake detectors: models trained on current datasets lose 43% performance when tested on P$^{2}$V, with performance measured as th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.10949","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/2508.10949/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":"2508.10949","created_at":"2026-07-05T11:54:16.047705+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.10949v1","created_at":"2026-07-05T11:54:16.047705+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.10949","created_at":"2026-07-05T11:54:16.047705+00:00"},{"alias_kind":"pith_short_12","alias_value":"AXS4FXFNIBAE","created_at":"2026-07-05T11:54:16.047705+00:00"},{"alias_kind":"pith_short_16","alias_value":"AXS4FXFNIBAEOZ7K","created_at":"2026-07-05T11:54:16.047705+00:00"},{"alias_kind":"pith_short_8","alias_value":"AXS4FXFN","created_at":"2026-07-05T11:54:16.047705+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.09568","citing_title":"RADAR Challenge 2026: Robust Audio Deepfake Recognition under Media Transformations","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09568","citing_title":"RADAR Challenge 2026: Robust Audio Deepfake Recognition under Media Transformations","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06035","citing_title":"Quantum Kernels for Audio Deepfake Detection Using Spectrogram Patch Features","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW","json":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW.json","graph_json":"https://pith.science/api/pith-number/AXS4FXFNIBAEOZ7KGYGAOU2XOW/graph.json","events_json":"https://pith.science/api/pith-number/AXS4FXFNIBAEOZ7KGYGAOU2XOW/events.json","paper":"https://pith.science/paper/AXS4FXFN"},"agent_actions":{"view_html":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW","download_json":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW.json","view_paper":"https://pith.science/paper/AXS4FXFN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.10949&json=true","fetch_graph":"https://pith.science/api/pith-number/AXS4FXFNIBAEOZ7KGYGAOU2XOW/graph.json","fetch_events":"https://pith.science/api/pith-number/AXS4FXFNIBAEOZ7KGYGAOU2XOW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW/action/storage_attestation","attest_author":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW/action/author_attestation","sign_citation":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW/action/citation_signature","submit_replication":"https://pith.science/pith/AXS4FXFNIBAEOZ7KGYGAOU2XOW/action/replication_record"}},"created_at":"2026-07-05T11:54:16.047705+00:00","updated_at":"2026-07-05T11:54:16.047705+00:00"}