{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IX7VAFWQOQXFFDZ55M6NSUDZC5","short_pith_number":"pith:IX7VAFWQ","schema_version":"1.0","canonical_sha256":"45ff5016d0742e528f3deb3cd950791776f2b91ccae5031c0f4a7bc1d155bc7c","source":{"kind":"arxiv","id":"2401.18045","version":1},"attestation_state":"computed","paper":{"title":"SpeechComposer: Unifying Multiple Speech Tasks with Prompt Composition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Chenda Li, Ruihua Song, Shinji Watanabe, Soumi Maiti, Wangyou Zhang, Xihua Wang, Yifan Peng, Yihan Wu, Yuyue Wang","submitted_at":"2024-01-31T18:06:29Z","abstract_excerpt":"Recent advancements in language models have significantly enhanced performance in multiple speech-related tasks. Existing speech language models typically utilize task-dependent prompt tokens to unify various speech tasks in a single model. However, this design omits the intrinsic connections between different speech tasks, which can potentially boost the performance of each task. In this work, we propose a novel decoder-only speech language model, SpeechComposer, that can unify common speech tasks by composing a fixed set of prompt tokens. Built upon four primary tasks -- speech synthesis, sp"},"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":"2401.18045","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-31T18:06:29Z","cross_cats_sorted":["cs.AI","cs.SD","eess.AS"],"title_canon_sha256":"4f0cae829d7fdbbb7100ab55a8690c1ff11ae277229df2f4cbc7d606314774a7","abstract_canon_sha256":"8518df9b0f36d2195ba33c340dcd502616ed277b85834aa8d2bb18864f79371d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:39:46.170484Z","signature_b64":"pf1W9D9tAXBOD6PYob992FvH8Mty7cv4rEFpa/EDn4zPf/lZm/8EamliU7WGGt03dajXqDu7M5HWufVqkHUqAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"45ff5016d0742e528f3deb3cd950791776f2b91ccae5031c0f4a7bc1d155bc7c","last_reissued_at":"2026-07-05T07:39:46.170082Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:39:46.170082Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SpeechComposer: Unifying Multiple Speech Tasks with Prompt Composition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Chenda Li, Ruihua Song, Shinji Watanabe, Soumi Maiti, Wangyou Zhang, Xihua Wang, Yifan Peng, Yihan Wu, Yuyue Wang","submitted_at":"2024-01-31T18:06:29Z","abstract_excerpt":"Recent advancements in language models have significantly enhanced performance in multiple speech-related tasks. Existing speech language models typically utilize task-dependent prompt tokens to unify various speech tasks in a single model. However, this design omits the intrinsic connections between different speech tasks, which can potentially boost the performance of each task. In this work, we propose a novel decoder-only speech language model, SpeechComposer, that can unify common speech tasks by composing a fixed set of prompt tokens. Built upon four primary tasks -- speech synthesis, sp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.18045","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/2401.18045/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":"2401.18045","created_at":"2026-07-05T07:39:46.170141+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.18045v1","created_at":"2026-07-05T07:39:46.170141+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.18045","created_at":"2026-07-05T07:39:46.170141+00:00"},{"alias_kind":"pith_short_12","alias_value":"IX7VAFWQOQXF","created_at":"2026-07-05T07:39:46.170141+00:00"},{"alias_kind":"pith_short_16","alias_value":"IX7VAFWQOQXFFDZ5","created_at":"2026-07-05T07:39:46.170141+00:00"},{"alias_kind":"pith_short_8","alias_value":"IX7VAFWQ","created_at":"2026-07-05T07:39:46.170141+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.19856","citing_title":"TokenVerse++: Towards Flexible Multitask Learning with Dynamic Task Activation","ref_index":4,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5","json":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5.json","graph_json":"https://pith.science/api/pith-number/IX7VAFWQOQXFFDZ55M6NSUDZC5/graph.json","events_json":"https://pith.science/api/pith-number/IX7VAFWQOQXFFDZ55M6NSUDZC5/events.json","paper":"https://pith.science/paper/IX7VAFWQ"},"agent_actions":{"view_html":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5","download_json":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5.json","view_paper":"https://pith.science/paper/IX7VAFWQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.18045&json=true","fetch_graph":"https://pith.science/api/pith-number/IX7VAFWQOQXFFDZ55M6NSUDZC5/graph.json","fetch_events":"https://pith.science/api/pith-number/IX7VAFWQOQXFFDZ55M6NSUDZC5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5/action/storage_attestation","attest_author":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5/action/author_attestation","sign_citation":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5/action/citation_signature","submit_replication":"https://pith.science/pith/IX7VAFWQOQXFFDZ55M6NSUDZC5/action/replication_record"}},"created_at":"2026-07-05T07:39:46.170141+00:00","updated_at":"2026-07-05T07:39:46.170141+00:00"}