{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TFBRISTNB7XN2P672PRXVXK5XY","short_pith_number":"pith:TFBRISTN","schema_version":"1.0","canonical_sha256":"9943144a6d0feedd3fdfd3e37add5dbe27a8cc6f88b18c6ec4c48acbd37a3f43","source":{"kind":"arxiv","id":"2603.13952","version":3},"attestation_state":"computed","paper":{"title":"LLM-Guided Reinforcement Learning for Audio-Visual Speech Enhancement","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","eess.AS"],"primary_cat":"cs.SD","authors_text":"Chih-Ning Chen, Fan-Gang Zeng, Hsin-Min Wang, Jen-Cheng Hou, Shao-Yi Chien, Yu Tsao","submitted_at":"2026-03-14T14:01:45Z","abstract_excerpt":"In existing Audio-Visual Speech Enhancement (AVSE) methods, objectives such as Scale-Invariant Signal-to-Noise Ratio (SI-SNR) and Mean Squared Error (MSE) are widely used; however, their correlation with perceived speech quality is often suboptimal and provides limited interpretability for optimization. This work proposes a reinforcement learning-based AVSE framework with a Large Language Model (LLM)-based interpretable reward model. An audio LLM generates natural language descriptions of enhanced speech, which are converted by a sentiment analysis model into a 1-5 rating score serving as the "},"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":"2603.13952","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2026-03-14T14:01:45Z","cross_cats_sorted":["cs.AI","eess.AS"],"title_canon_sha256":"76a702a3269c2ef2055b0d5fb3d43ceeb44ea589bef353d0a7f6a4f0aeaedc6e","abstract_canon_sha256":"5a85345cf8037da4c1328bf58f7ee0cdcd7a73fc5739c5ced93df23939c299ef"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-16T01:22:37.384614Z","signature_b64":"Utu8jsH+kCvXsrsfcyl7Bf+yX8xXQocZvRCkkx+tsQ9PYgWQC8npKYbq0TcCy+PADJPis1rEv/ttAtAbULENBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9943144a6d0feedd3fdfd3e37add5dbe27a8cc6f88b18c6ec4c48acbd37a3f43","last_reissued_at":"2026-07-16T01:22:37.383715Z","signature_status":"signed_v1","first_computed_at":"2026-07-16T01:22:37.383715Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLM-Guided Reinforcement Learning for Audio-Visual Speech Enhancement","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","eess.AS"],"primary_cat":"cs.SD","authors_text":"Chih-Ning Chen, Fan-Gang Zeng, Hsin-Min Wang, Jen-Cheng Hou, Shao-Yi Chien, Yu Tsao","submitted_at":"2026-03-14T14:01:45Z","abstract_excerpt":"In existing Audio-Visual Speech Enhancement (AVSE) methods, objectives such as Scale-Invariant Signal-to-Noise Ratio (SI-SNR) and Mean Squared Error (MSE) are widely used; however, their correlation with perceived speech quality is often suboptimal and provides limited interpretability for optimization. This work proposes a reinforcement learning-based AVSE framework with a Large Language Model (LLM)-based interpretable reward model. An audio LLM generates natural language descriptions of enhanced speech, which are converted by a sentiment analysis model into a 1-5 rating score serving as the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.13952","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/2603.13952/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":"2603.13952","created_at":"2026-07-16T01:22:37.384147+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.13952v3","created_at":"2026-07-16T01:22:37.384147+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.13952","created_at":"2026-07-16T01:22:37.384147+00:00"},{"alias_kind":"pith_short_12","alias_value":"TFBRISTNB7XN","created_at":"2026-07-16T01:22:37.384147+00:00"},{"alias_kind":"pith_short_16","alias_value":"TFBRISTNB7XN2P67","created_at":"2026-07-16T01:22:37.384147+00:00"},{"alias_kind":"pith_short_8","alias_value":"TFBRISTN","created_at":"2026-07-16T01:22:37.384147+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/TFBRISTNB7XN2P672PRXVXK5XY","json":"https://pith.science/pith/TFBRISTNB7XN2P672PRXVXK5XY.json","graph_json":"https://pith.science/api/pith-number/TFBRISTNB7XN2P672PRXVXK5XY/graph.json","events_json":"https://pith.science/api/pith-number/TFBRISTNB7XN2P672PRXVXK5XY/events.json","paper":"https://pith.science/paper/TFBRISTN"},"agent_actions":{"view_html":"https://pith.science/pith/TFBRISTNB7XN2P672PRXVXK5XY","download_json":"https://pith.science/pith/TFBRISTNB7XN2P672PRXVXK5XY.json","view_paper":"https://pith.science/paper/TFBRISTN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.13952&json=true","fetch_graph":"https://pith.science/api/pith-number/TFBRISTNB7XN2P672PRXVXK5XY/graph.json","fetch_events":"https://pith.science/api/pith-number/TFBRISTNB7XN2P672PRXVXK5XY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TFBRISTNB7XN2P672PRXVXK5XY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TFBRISTNB7XN2P672PRXVXK5XY/action/storage_attestation","attest_author":"https://pith.science/pith/TFBRISTNB7XN2P672PRXVXK5XY/action/author_attestation","sign_citation":"https://pith.science/pith/TFBRISTNB7XN2P672PRXVXK5XY/action/citation_signature","submit_replication":"https://pith.science/pith/TFBRISTNB7XN2P672PRXVXK5XY/action/replication_record"}},"created_at":"2026-07-16T01:22:37.384147+00:00","updated_at":"2026-07-16T01:22:37.384147+00:00"}