{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:UJCCHBZOGM7VPHXWSUXL3FQVER","short_pith_number":"pith:UJCCHBZO","schema_version":"1.0","canonical_sha256":"a24423872e333f579ef6952ebd96152463bd9afc95b1fc6c04470d34990c5956","source":{"kind":"arxiv","id":"2310.05224","version":1},"attestation_state":"computed","paper":{"title":"Generative Spoken Language Model based on continuous word-sized audio tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Benoit Sagot, Emmanuel Dupoux, Gabriel Synnaeve, Jade Copet, Robin Algayres, Tu Anh Nguyen, Yossi Adi","submitted_at":"2023-10-08T16:46:14Z","abstract_excerpt":"In NLP, text language models based on words or subwords are known to outperform their character-based counterparts. Yet, in the speech community, the standard input of spoken LMs are 20ms or 40ms-long discrete units (shorter than a phoneme). Taking inspiration from word-based LM, we introduce a Generative Spoken Language Model (GSLM) based on word-size continuous-valued audio embeddings that can generate diverse and expressive language output. This is obtained by replacing lookup table for lexical types with a Lexical Embedding function, the cross entropy loss by a contrastive loss, and multin"},"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":"2310.05224","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-08T16:46:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e4982d6278f864d661a3d0263fb2910328b92221a26bb1f134e705cb1a902fa2","abstract_canon_sha256":"ed515d93a8d637a8aeafcf9a5844cb55ac30897edb840ae44ec7ba2e8bbde5b6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:58:28.329130Z","signature_b64":"l956hPKkrXDmfus8l79wSviQAxTP1dkJT5vKkgUtmQJFRxmv7YdBNUV5zBeo7DN12Yrid05pODgWaq0v1jMgDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a24423872e333f579ef6952ebd96152463bd9afc95b1fc6c04470d34990c5956","last_reissued_at":"2026-07-05T06:58:28.328620Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:58:28.328620Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generative Spoken Language Model based on continuous word-sized audio tokens","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Benoit Sagot, Emmanuel Dupoux, Gabriel Synnaeve, Jade Copet, Robin Algayres, Tu Anh Nguyen, Yossi Adi","submitted_at":"2023-10-08T16:46:14Z","abstract_excerpt":"In NLP, text language models based on words or subwords are known to outperform their character-based counterparts. Yet, in the speech community, the standard input of spoken LMs are 20ms or 40ms-long discrete units (shorter than a phoneme). Taking inspiration from word-based LM, we introduce a Generative Spoken Language Model (GSLM) based on word-size continuous-valued audio embeddings that can generate diverse and expressive language output. This is obtained by replacing lookup table for lexical types with a Lexical Embedding function, the cross entropy loss by a contrastive loss, and multin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.05224","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/2310.05224/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":"2310.05224","created_at":"2026-07-05T06:58:28.328705+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.05224v1","created_at":"2026-07-05T06:58:28.328705+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.05224","created_at":"2026-07-05T06:58:28.328705+00:00"},{"alias_kind":"pith_short_12","alias_value":"UJCCHBZOGM7V","created_at":"2026-07-05T06:58:28.328705+00:00"},{"alias_kind":"pith_short_16","alias_value":"UJCCHBZOGM7VPHXW","created_at":"2026-07-05T06:58:28.328705+00:00"},{"alias_kind":"pith_short_8","alias_value":"UJCCHBZO","created_at":"2026-07-05T06:58:28.328705+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/UJCCHBZOGM7VPHXWSUXL3FQVER","json":"https://pith.science/pith/UJCCHBZOGM7VPHXWSUXL3FQVER.json","graph_json":"https://pith.science/api/pith-number/UJCCHBZOGM7VPHXWSUXL3FQVER/graph.json","events_json":"https://pith.science/api/pith-number/UJCCHBZOGM7VPHXWSUXL3FQVER/events.json","paper":"https://pith.science/paper/UJCCHBZO"},"agent_actions":{"view_html":"https://pith.science/pith/UJCCHBZOGM7VPHXWSUXL3FQVER","download_json":"https://pith.science/pith/UJCCHBZOGM7VPHXWSUXL3FQVER.json","view_paper":"https://pith.science/paper/UJCCHBZO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.05224&json=true","fetch_graph":"https://pith.science/api/pith-number/UJCCHBZOGM7VPHXWSUXL3FQVER/graph.json","fetch_events":"https://pith.science/api/pith-number/UJCCHBZOGM7VPHXWSUXL3FQVER/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UJCCHBZOGM7VPHXWSUXL3FQVER/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UJCCHBZOGM7VPHXWSUXL3FQVER/action/storage_attestation","attest_author":"https://pith.science/pith/UJCCHBZOGM7VPHXWSUXL3FQVER/action/author_attestation","sign_citation":"https://pith.science/pith/UJCCHBZOGM7VPHXWSUXL3FQVER/action/citation_signature","submit_replication":"https://pith.science/pith/UJCCHBZOGM7VPHXWSUXL3FQVER/action/replication_record"}},"created_at":"2026-07-05T06:58:28.328705+00:00","updated_at":"2026-07-05T06:58:28.328705+00:00"}