{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BVCAATDFF6OQNL6NCYS4Q4FHSE","short_pith_number":"pith:BVCAATDF","schema_version":"1.0","canonical_sha256":"0d44004c652f9d06afcd1625c870a79133b57a4f6816aa2ef7fe5c1c723a8e0a","source":{"kind":"arxiv","id":"2508.05011","version":1},"attestation_state":"computed","paper":{"title":"Towards Hallucination-Free Music: A Reinforcement Learning Preference Optimization Framework for Reliable Song Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","eess.AS"],"primary_cat":"cs.SD","authors_text":"Chenyu Yang, Dong Yu, Guangzheng Li, Hangting Chen, Huaicheng Zhang, Qijun Huang, Shuai Wang, Shun Lei, Wei Tan, Yixuan Zhang, Zhiyong Wu","submitted_at":"2025-08-07T03:49:18Z","abstract_excerpt":"Recent advances in audio-based generative language models have accelerated AI-driven lyric-to-song generation. However, these models frequently suffer from content hallucination, producing outputs misaligned with the input lyrics and undermining musical coherence. Current supervised fine-tuning (SFT) approaches, limited by passive label-fitting, exhibit constrained self-improvement and poor hallucination mitigation. To address this core challenge, we propose a novel reinforcement learning (RL) framework leveraging preference optimization for hallucination control. Our key contributions include"},"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":"2508.05011","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SD","submitted_at":"2025-08-07T03:49:18Z","cross_cats_sorted":["cs.AI","eess.AS"],"title_canon_sha256":"10b4d2f42696925a8286920022bcacf076435e79be13f719fa42271db2420eb8","abstract_canon_sha256":"bcb1782faae7ebcf955a94439e156185691b8cacfb11937363c98a630508aa9e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:50:05.641963Z","signature_b64":"SkvT6yPIfLo8Zerdwr6cCkHUbLooCrMSZZWAU1SMZUj1YkPf1FC7+iV/mLJvIRsmwqJyWt/iJ8iNalDTyx57Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0d44004c652f9d06afcd1625c870a79133b57a4f6816aa2ef7fe5c1c723a8e0a","last_reissued_at":"2026-07-05T11:50:05.641486Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:50:05.641486Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Hallucination-Free Music: A Reinforcement Learning Preference Optimization Framework for Reliable Song Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","eess.AS"],"primary_cat":"cs.SD","authors_text":"Chenyu Yang, Dong Yu, Guangzheng Li, Hangting Chen, Huaicheng Zhang, Qijun Huang, Shuai Wang, Shun Lei, Wei Tan, Yixuan Zhang, Zhiyong Wu","submitted_at":"2025-08-07T03:49:18Z","abstract_excerpt":"Recent advances in audio-based generative language models have accelerated AI-driven lyric-to-song generation. However, these models frequently suffer from content hallucination, producing outputs misaligned with the input lyrics and undermining musical coherence. Current supervised fine-tuning (SFT) approaches, limited by passive label-fitting, exhibit constrained self-improvement and poor hallucination mitigation. To address this core challenge, we propose a novel reinforcement learning (RL) framework leveraging preference optimization for hallucination control. Our key contributions include"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.05011","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/2508.05011/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":"2508.05011","created_at":"2026-07-05T11:50:05.641546+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.05011v1","created_at":"2026-07-05T11:50:05.641546+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.05011","created_at":"2026-07-05T11:50:05.641546+00:00"},{"alias_kind":"pith_short_12","alias_value":"BVCAATDFF6OQ","created_at":"2026-07-05T11:50:05.641546+00:00"},{"alias_kind":"pith_short_16","alias_value":"BVCAATDFF6OQNL6N","created_at":"2026-07-05T11:50:05.641546+00:00"},{"alias_kind":"pith_short_8","alias_value":"BVCAATDF","created_at":"2026-07-05T11:50:05.641546+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE","json":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE.json","graph_json":"https://pith.science/api/pith-number/BVCAATDFF6OQNL6NCYS4Q4FHSE/graph.json","events_json":"https://pith.science/api/pith-number/BVCAATDFF6OQNL6NCYS4Q4FHSE/events.json","paper":"https://pith.science/paper/BVCAATDF"},"agent_actions":{"view_html":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE","download_json":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE.json","view_paper":"https://pith.science/paper/BVCAATDF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.05011&json=true","fetch_graph":"https://pith.science/api/pith-number/BVCAATDFF6OQNL6NCYS4Q4FHSE/graph.json","fetch_events":"https://pith.science/api/pith-number/BVCAATDFF6OQNL6NCYS4Q4FHSE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE/action/storage_attestation","attest_author":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE/action/author_attestation","sign_citation":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE/action/citation_signature","submit_replication":"https://pith.science/pith/BVCAATDFF6OQNL6NCYS4Q4FHSE/action/replication_record"}},"created_at":"2026-07-05T11:50:05.641546+00:00","updated_at":"2026-07-05T11:50:05.641546+00:00"}