{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:7SKGRJSYWACQV7HJRA5352LH3W","short_pith_number":"pith:7SKGRJSY","schema_version":"1.0","canonical_sha256":"fc9468a658b0050afce9883bbee967dd80a79ed6e241a6bd4844795f6ba97172","source":{"kind":"arxiv","id":"2401.17264","version":2},"attestation_state":"computed","paper":{"title":"Proactive Detection of Voice Cloning with Localized Watermarking","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.SD","authors_text":"Alexandre D\\'efossez, Hady Elsahar, Pierre Fernandez, Robin San Roman, Teddy Furon, Tuan Tran","submitted_at":"2024-01-30T18:56:22Z","abstract_excerpt":"In the rapidly evolving field of speech generative models, there is a pressing need to ensure audio authenticity against the risks of voice cloning. We present AudioSeal, the first audio watermarking technique designed specifically for localized detection of AI-generated speech. AudioSeal employs a generator/detector architecture trained jointly with a localization loss to enable localized watermark detection up to the sample level, and a novel perceptual loss inspired by auditory masking, that enables AudioSeal to achieve better imperceptibility. AudioSeal achieves state-of-the-art performanc"},"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":"2401.17264","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2024-01-30T18:56:22Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"bf752754ff81fd81f68547b2f9a09d56a76a32156765aad872ee56d59c81b93b","abstract_canon_sha256":"7f705eb350047b12a5ee1bc252a09d8a37378ad2c3012232575d1b8241323b08"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:26.905396Z","signature_b64":"vbDwJFlt247j8DXzs8QNG8yVxN6jFw7D85K8J4WBlem0jUBZtUCT0t3v2N5qLPKByFr9zcwx3VvjUPEKRRpBAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fc9468a658b0050afce9883bbee967dd80a79ed6e241a6bd4844795f6ba97172","last_reissued_at":"2026-07-05T08:28:26.904960Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:26.904960Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Proactive Detection of Voice Cloning with Localized Watermarking","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.SD","authors_text":"Alexandre D\\'efossez, Hady Elsahar, Pierre Fernandez, Robin San Roman, Teddy Furon, Tuan Tran","submitted_at":"2024-01-30T18:56:22Z","abstract_excerpt":"In the rapidly evolving field of speech generative models, there is a pressing need to ensure audio authenticity against the risks of voice cloning. We present AudioSeal, the first audio watermarking technique designed specifically for localized detection of AI-generated speech. AudioSeal employs a generator/detector architecture trained jointly with a localization loss to enable localized watermark detection up to the sample level, and a novel perceptual loss inspired by auditory masking, that enables AudioSeal to achieve better imperceptibility. AudioSeal achieves state-of-the-art performanc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.17264","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/2401.17264/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":"2401.17264","created_at":"2026-07-05T08:28:26.905015+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.17264v2","created_at":"2026-07-05T08:28:26.905015+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.17264","created_at":"2026-07-05T08:28:26.905015+00:00"},{"alias_kind":"pith_short_12","alias_value":"7SKGRJSYWACQ","created_at":"2026-07-05T08:28:26.905015+00:00"},{"alias_kind":"pith_short_16","alias_value":"7SKGRJSYWACQV7HJ","created_at":"2026-07-05T08:28:26.905015+00:00"},{"alias_kind":"pith_short_8","alias_value":"7SKGRJSY","created_at":"2026-07-05T08:28:26.905015+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23335","citing_title":"The Watermark Shortcut: How Provenance Marking Sabotages Audio Deepfake Detection","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30614","citing_title":"Audio Pirates: Black-box Audio Watermark Removal via Diffusion Priors","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01515","citing_title":"MelShield: Robust Mel-Domain Audio Watermarking for Provenance Attribution of AI Generated Synthesized Speech","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13776","citing_title":"Who Gets Flagged? The Pluralistic Evaluation Gap in AI Content Watermarking","ref_index":48,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W","json":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W.json","graph_json":"https://pith.science/api/pith-number/7SKGRJSYWACQV7HJRA5352LH3W/graph.json","events_json":"https://pith.science/api/pith-number/7SKGRJSYWACQV7HJRA5352LH3W/events.json","paper":"https://pith.science/paper/7SKGRJSY"},"agent_actions":{"view_html":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W","download_json":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W.json","view_paper":"https://pith.science/paper/7SKGRJSY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.17264&json=true","fetch_graph":"https://pith.science/api/pith-number/7SKGRJSYWACQV7HJRA5352LH3W/graph.json","fetch_events":"https://pith.science/api/pith-number/7SKGRJSYWACQV7HJRA5352LH3W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W/action/storage_attestation","attest_author":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W/action/author_attestation","sign_citation":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W/action/citation_signature","submit_replication":"https://pith.science/pith/7SKGRJSYWACQV7HJRA5352LH3W/action/replication_record"}},"created_at":"2026-07-05T08:28:26.905015+00:00","updated_at":"2026-07-05T08:28:26.905015+00:00"}