{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:2CEISFTFVC6P6DFGAOMNT3VF7L","short_pith_number":"pith:2CEISFTF","schema_version":"1.0","canonical_sha256":"d088891665a8bcff0ca60398d9eea5faf09a32cdd830210a653a43d594e9e553","source":{"kind":"arxiv","id":"2212.04559","version":1},"attestation_state":"computed","paper":{"title":"SpeechLMScore: Evaluating speech generation using speech language model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Shinji Watanabe, Soumi Maiti, Takaaki Saeki, Yifan Peng","submitted_at":"2022-12-08T21:00:15Z","abstract_excerpt":"While human evaluation is the most reliable metric for evaluating speech generation systems, it is generally costly and time-consuming. Previous studies on automatic speech quality assessment address the problem by predicting human evaluation scores with machine learning models. However, they rely on supervised learning and thus suffer from high annotation costs and domain-shift problems. We propose SpeechLMScore, an unsupervised metric to evaluate generated speech using a speech-language model. SpeechLMScore computes the average log-probability of a speech signal by mapping it into discrete t"},"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":"2212.04559","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2022-12-08T21:00:15Z","cross_cats_sorted":["cs.LG","cs.SD"],"title_canon_sha256":"441f0f6bbde1ea271f91a088d017624239e41f51e0489e59f1df37fdc4225371","abstract_canon_sha256":"43c4fb8e2b9c6fb7b31caff3532fa482603139c9bbabb9613f51f3b42c926ae2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:23:49.071429Z","signature_b64":"CwXT7/4ZXPUiZZoeIBmU+0CVGHMv3s9ElNp9L83jFHVn1RGm9zAKdWo+IukatkhKhQtJjE9cu8X1tPJGjqIJBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d088891665a8bcff0ca60398d9eea5faf09a32cdd830210a653a43d594e9e553","last_reissued_at":"2026-07-05T05:23:49.070997Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:23:49.070997Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SpeechLMScore: Evaluating speech generation using speech language model","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Shinji Watanabe, Soumi Maiti, Takaaki Saeki, Yifan Peng","submitted_at":"2022-12-08T21:00:15Z","abstract_excerpt":"While human evaluation is the most reliable metric for evaluating speech generation systems, it is generally costly and time-consuming. Previous studies on automatic speech quality assessment address the problem by predicting human evaluation scores with machine learning models. However, they rely on supervised learning and thus suffer from high annotation costs and domain-shift problems. We propose SpeechLMScore, an unsupervised metric to evaluate generated speech using a speech-language model. SpeechLMScore computes the average log-probability of a speech signal by mapping it into discrete t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.04559","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/2212.04559/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":"2212.04559","created_at":"2026-07-05T05:23:49.071049+00:00"},{"alias_kind":"arxiv_version","alias_value":"2212.04559v1","created_at":"2026-07-05T05:23:49.071049+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.04559","created_at":"2026-07-05T05:23:49.071049+00:00"},{"alias_kind":"pith_short_12","alias_value":"2CEISFTFVC6P","created_at":"2026-07-05T05:23:49.071049+00:00"},{"alias_kind":"pith_short_16","alias_value":"2CEISFTFVC6P6DFG","created_at":"2026-07-05T05:23:49.071049+00:00"},{"alias_kind":"pith_short_8","alias_value":"2CEISFTF","created_at":"2026-07-05T05:23:49.071049+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.21156","citing_title":"Model as Loss: A Self-Consistent Training Paradigm","ref_index":2022,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L","json":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L.json","graph_json":"https://pith.science/api/pith-number/2CEISFTFVC6P6DFGAOMNT3VF7L/graph.json","events_json":"https://pith.science/api/pith-number/2CEISFTFVC6P6DFGAOMNT3VF7L/events.json","paper":"https://pith.science/paper/2CEISFTF"},"agent_actions":{"view_html":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L","download_json":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L.json","view_paper":"https://pith.science/paper/2CEISFTF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2212.04559&json=true","fetch_graph":"https://pith.science/api/pith-number/2CEISFTFVC6P6DFGAOMNT3VF7L/graph.json","fetch_events":"https://pith.science/api/pith-number/2CEISFTFVC6P6DFGAOMNT3VF7L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L/action/storage_attestation","attest_author":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L/action/author_attestation","sign_citation":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L/action/citation_signature","submit_replication":"https://pith.science/pith/2CEISFTFVC6P6DFGAOMNT3VF7L/action/replication_record"}},"created_at":"2026-07-05T05:23:49.071049+00:00","updated_at":"2026-07-05T05:23:49.071049+00:00"}