{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CMNEBIVH4KZAKG75VE75CPIRO5","short_pith_number":"pith:CMNEBIVH","schema_version":"1.0","canonical_sha256":"131a40a2a7e2b2051bfda93fd13d1177549cde90ff3a1d6ad776ffc8929581dd","source":{"kind":"arxiv","id":"2502.03128","version":1},"attestation_state":"computed","paper":{"title":"Metis: A Foundation Speech Generation Model with Masked Generative Pre-training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","eess.AS","eess.SP"],"primary_cat":"cs.SD","authors_text":"Huan Liao, Jiachen Zheng, Junan Zhang, Xueyao Zhang, Yuancheng Wang, Zhizheng Wu","submitted_at":"2025-02-05T12:36:21Z","abstract_excerpt":"We introduce Metis, a foundation model for unified speech generation. Unlike previous task-specific or multi-task models, Metis follows a pre-training and fine-tuning paradigm. It is pre-trained on large-scale unlabeled speech data using masked generative modeling and then fine-tuned to adapt to diverse speech generation tasks. Specifically, 1) Metis utilizes two discrete speech representations: SSL tokens derived from speech self-supervised learning (SSL) features, and acoustic tokens directly quantized from waveforms. 2) Metis performs masked generative pre-training on SSL tokens, utilizing "},"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":"2502.03128","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SD","submitted_at":"2025-02-05T12:36:21Z","cross_cats_sorted":["cs.AI","cs.LG","eess.AS","eess.SP"],"title_canon_sha256":"b78d0f8361b7e5a01051d4d30533094faebf0c08d5d6b481c216c0a4461412ae","abstract_canon_sha256":"715068f9159b8243edd6cad071cbe3ea3fca6719ac7dff08d9310336c529b573"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:10:00.179129Z","signature_b64":"kIuKsei7WewauJNuo5fD9lnlkEwakwQM3lBu0MPOx8QLnIcxVy5sSmMC086VgIhuj0/H3GhCCDa1Si/J81+CBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"131a40a2a7e2b2051bfda93fd13d1177549cde90ff3a1d6ad776ffc8929581dd","last_reissued_at":"2026-07-05T10:10:00.178650Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:10:00.178650Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Metis: A Foundation Speech Generation Model with Masked Generative Pre-training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","eess.AS","eess.SP"],"primary_cat":"cs.SD","authors_text":"Huan Liao, Jiachen Zheng, Junan Zhang, Xueyao Zhang, Yuancheng Wang, Zhizheng Wu","submitted_at":"2025-02-05T12:36:21Z","abstract_excerpt":"We introduce Metis, a foundation model for unified speech generation. Unlike previous task-specific or multi-task models, Metis follows a pre-training and fine-tuning paradigm. It is pre-trained on large-scale unlabeled speech data using masked generative modeling and then fine-tuned to adapt to diverse speech generation tasks. Specifically, 1) Metis utilizes two discrete speech representations: SSL tokens derived from speech self-supervised learning (SSL) features, and acoustic tokens directly quantized from waveforms. 2) Metis performs masked generative pre-training on SSL tokens, utilizing "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.03128","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/2502.03128/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":"2502.03128","created_at":"2026-07-05T10:10:00.178707+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.03128v1","created_at":"2026-07-05T10:10:00.178707+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.03128","created_at":"2026-07-05T10:10:00.178707+00:00"},{"alias_kind":"pith_short_12","alias_value":"CMNEBIVH4KZA","created_at":"2026-07-05T10:10:00.178707+00:00"},{"alias_kind":"pith_short_16","alias_value":"CMNEBIVH4KZAKG75","created_at":"2026-07-05T10:10:00.178707+00:00"},{"alias_kind":"pith_short_8","alias_value":"CMNEBIVH","created_at":"2026-07-05T10:10:00.178707+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18323","citing_title":"Reliable Neural-Codec Text-to-Speech by ASR Self-Verification and Distillation: Near-Zero Catastrophic Failures Across Models and Codecs","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31247","citing_title":"FlexiSLM: A Dynamic and Controllable Frame Rate Spoken Language Model","ref_index":203,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5","json":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5.json","graph_json":"https://pith.science/api/pith-number/CMNEBIVH4KZAKG75VE75CPIRO5/graph.json","events_json":"https://pith.science/api/pith-number/CMNEBIVH4KZAKG75VE75CPIRO5/events.json","paper":"https://pith.science/paper/CMNEBIVH"},"agent_actions":{"view_html":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5","download_json":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5.json","view_paper":"https://pith.science/paper/CMNEBIVH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.03128&json=true","fetch_graph":"https://pith.science/api/pith-number/CMNEBIVH4KZAKG75VE75CPIRO5/graph.json","fetch_events":"https://pith.science/api/pith-number/CMNEBIVH4KZAKG75VE75CPIRO5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5/action/storage_attestation","attest_author":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5/action/author_attestation","sign_citation":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5/action/citation_signature","submit_replication":"https://pith.science/pith/CMNEBIVH4KZAKG75VE75CPIRO5/action/replication_record"}},"created_at":"2026-07-05T10:10:00.178707+00:00","updated_at":"2026-07-05T10:10:00.178707+00:00"}