{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ADSSIOZNYONMDAL5K5Y3PH2XPT","short_pith_number":"pith:ADSSIOZN","schema_version":"1.0","canonical_sha256":"00e5243b2dc39ac1817d5771b79f577cf8f0c02e165b3046114ee244098ff84e","source":{"kind":"arxiv","id":"2504.10352","version":3},"attestation_state":"computed","paper":{"title":"Pseudo-Autoregressive Neural Codec Language Models for Efficient Zero-Shot Text-to-Speech Synthesis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"eess.AS","authors_text":"Haibin Wu, Haiyang Sun, Hui Wang, Jianwei Yu, Jinyu Li, Kai Yu, Lingwei Meng, Shujie Liu, Xie Chen, Yan Lu, Yanqing Liu, Yifan Yang, Yuxuan Hu","submitted_at":"2025-04-14T16:03:21Z","abstract_excerpt":"Recent zero-shot text-to-speech (TTS) systems face a common dilemma: autoregressive (AR) models suffer from slow generation and lack duration controllability, while non-autoregressive (NAR) models lack temporal modeling and typically require complex designs. In this paper, we introduce a novel pseudo-autoregressive (PAR) codec language modeling approach that unifies AR and NAR modeling. Combining explicit temporal modeling from AR with parallel generation from NAR, PAR generates dynamic-length spans at fixed time steps. Building on PAR, we propose PALLE, a two-stage TTS system that leverages P"},"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":"2504.10352","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2025-04-14T16:03:21Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"cb31c851b70abb9a4a56c4201b1cdc12b7290381e342f997c1131ed27db706fb","abstract_canon_sha256":"b252b594419b39cd758c8626cd23cf072ae249dedb742e1d82aa463785af20c2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:48:48.653858Z","signature_b64":"fkeu0q7igc+mk5ji74cfwbWexyZw1Ru5IEbQgER4w/j5ehhr0V6Q1YSsp8awvHbsy/YUQayJFvi/gxPlrRSJAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"00e5243b2dc39ac1817d5771b79f577cf8f0c02e165b3046114ee244098ff84e","last_reissued_at":"2026-07-05T11:48:48.653178Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:48:48.653178Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Pseudo-Autoregressive Neural Codec Language Models for Efficient Zero-Shot Text-to-Speech Synthesis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"eess.AS","authors_text":"Haibin Wu, Haiyang Sun, Hui Wang, Jianwei Yu, Jinyu Li, Kai Yu, Lingwei Meng, Shujie Liu, Xie Chen, Yan Lu, Yanqing Liu, Yifan Yang, Yuxuan Hu","submitted_at":"2025-04-14T16:03:21Z","abstract_excerpt":"Recent zero-shot text-to-speech (TTS) systems face a common dilemma: autoregressive (AR) models suffer from slow generation and lack duration controllability, while non-autoregressive (NAR) models lack temporal modeling and typically require complex designs. In this paper, we introduce a novel pseudo-autoregressive (PAR) codec language modeling approach that unifies AR and NAR modeling. Combining explicit temporal modeling from AR with parallel generation from NAR, PAR generates dynamic-length spans at fixed time steps. Building on PAR, we propose PALLE, a two-stage TTS system that leverages P"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.10352","kind":"arxiv","version":3},"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/2504.10352/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":"2504.10352","created_at":"2026-07-05T11:48:48.653273+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.10352v3","created_at":"2026-07-05T11:48:48.653273+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.10352","created_at":"2026-07-05T11:48:48.653273+00:00"},{"alias_kind":"pith_short_12","alias_value":"ADSSIOZNYONM","created_at":"2026-07-05T11:48:48.653273+00:00"},{"alias_kind":"pith_short_16","alias_value":"ADSSIOZNYONMDAL5","created_at":"2026-07-05T11:48:48.653273+00:00"},{"alias_kind":"pith_short_8","alias_value":"ADSSIOZN","created_at":"2026-07-05T11:48:48.653273+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08256","citing_title":"Best-of-$N$ TTS Evaluation is Confounded by ASR Family Alignment","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2606.31247","citing_title":"FlexiSLM: A Dynamic and Controllable Frame Rate Spoken Language Model","ref_index":233,"is_internal_anchor":false},{"citing_arxiv_id":"2507.09318","citing_title":"ZipVoice-Dialog: Non-Autoregressive Spoken Dialogue Generation with Flow Matching","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT","json":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT.json","graph_json":"https://pith.science/api/pith-number/ADSSIOZNYONMDAL5K5Y3PH2XPT/graph.json","events_json":"https://pith.science/api/pith-number/ADSSIOZNYONMDAL5K5Y3PH2XPT/events.json","paper":"https://pith.science/paper/ADSSIOZN"},"agent_actions":{"view_html":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT","download_json":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT.json","view_paper":"https://pith.science/paper/ADSSIOZN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.10352&json=true","fetch_graph":"https://pith.science/api/pith-number/ADSSIOZNYONMDAL5K5Y3PH2XPT/graph.json","fetch_events":"https://pith.science/api/pith-number/ADSSIOZNYONMDAL5K5Y3PH2XPT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT/action/storage_attestation","attest_author":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT/action/author_attestation","sign_citation":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT/action/citation_signature","submit_replication":"https://pith.science/pith/ADSSIOZNYONMDAL5K5Y3PH2XPT/action/replication_record"}},"created_at":"2026-07-05T11:48:48.653273+00:00","updated_at":"2026-07-05T11:48:48.653273+00:00"}