{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:EBKRLX26ILQ7L3KYP66RVVXOTA","short_pith_number":"pith:EBKRLX26","schema_version":"1.0","canonical_sha256":"205515df5e42e1f5ed587fbd1ad6ee982858370252e51f9b1980fb665519a1df","source":{"kind":"arxiv","id":"2508.17863","version":1},"attestation_state":"computed","paper":{"title":"Speech Discrete Tokens or Continuous Features? A Comparative Analysis for Spoken Language Understanding in SpeechLLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.SD"],"primary_cat":"cs.CL","authors_text":"Dingdong Wang, Dongchao Yang, Helen Meng, Junan Li, Mingyu Cui, Xueyuan Chen","submitted_at":"2025-08-25T10:16:07Z","abstract_excerpt":"With the rise of Speech Large Language Models (SpeechLLMs), two dominant approaches have emerged for speech processing: discrete tokens and continuous features. Each approach has demonstrated strong capabilities in audio-related processing tasks. However, the performance gap between these two paradigms has not been thoroughly explored. To address this gap, we present a fair comparison of self-supervised learning (SSL)-based discrete and continuous features under the same experimental settings. We evaluate their performance across six spoken language understanding-related tasks using both small"},"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":"2508.17863","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-25T10:16:07Z","cross_cats_sorted":["cs.SD"],"title_canon_sha256":"4c02ffe74ce4a1eac41cb0a9fa48d5149c561f3a1e90ee8854c52e72357c2340","abstract_canon_sha256":"55b07e4dab311c2ceed66175fd2882dc887bd1cd8713f88cf3f9fe8aca351744"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:58:49.269389Z","signature_b64":"ApL82MeZFz663VdZnbBk5JlZvJHGnOeUjKl7AmN5Zxv2plSfyQr69JrSGGWwoBVVD8T4SCQHfPla0/cabaTUAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"205515df5e42e1f5ed587fbd1ad6ee982858370252e51f9b1980fb665519a1df","last_reissued_at":"2026-07-05T11:58:49.268889Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:58:49.268889Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Speech Discrete Tokens or Continuous Features? A Comparative Analysis for Spoken Language Understanding in SpeechLLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.SD"],"primary_cat":"cs.CL","authors_text":"Dingdong Wang, Dongchao Yang, Helen Meng, Junan Li, Mingyu Cui, Xueyuan Chen","submitted_at":"2025-08-25T10:16:07Z","abstract_excerpt":"With the rise of Speech Large Language Models (SpeechLLMs), two dominant approaches have emerged for speech processing: discrete tokens and continuous features. Each approach has demonstrated strong capabilities in audio-related processing tasks. However, the performance gap between these two paradigms has not been thoroughly explored. To address this gap, we present a fair comparison of self-supervised learning (SSL)-based discrete and continuous features under the same experimental settings. We evaluate their performance across six spoken language understanding-related tasks using both small"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.17863","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/2508.17863/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":"2508.17863","created_at":"2026-07-05T11:58:49.268955+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.17863v1","created_at":"2026-07-05T11:58:49.268955+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.17863","created_at":"2026-07-05T11:58:49.268955+00:00"},{"alias_kind":"pith_short_12","alias_value":"EBKRLX26ILQ7","created_at":"2026-07-05T11:58:49.268955+00:00"},{"alias_kind":"pith_short_16","alias_value":"EBKRLX26ILQ7L3KY","created_at":"2026-07-05T11:58:49.268955+00:00"},{"alias_kind":"pith_short_8","alias_value":"EBKRLX26","created_at":"2026-07-05T11:58:49.268955+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10368","citing_title":"Speech Meets ELF: Audio Conditional Continuous-Target Diffusion for Speech Recognition and Translation","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27627","citing_title":"HybridCodec: Modeling Discrete and Continuous Representations for Efficient Speech Language Models","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2509.14804","citing_title":"Towards Building Speech Large Language Models for Multitask Understanding in Low-Resource Languages","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA","json":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA.json","graph_json":"https://pith.science/api/pith-number/EBKRLX26ILQ7L3KYP66RVVXOTA/graph.json","events_json":"https://pith.science/api/pith-number/EBKRLX26ILQ7L3KYP66RVVXOTA/events.json","paper":"https://pith.science/paper/EBKRLX26"},"agent_actions":{"view_html":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA","download_json":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA.json","view_paper":"https://pith.science/paper/EBKRLX26","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.17863&json=true","fetch_graph":"https://pith.science/api/pith-number/EBKRLX26ILQ7L3KYP66RVVXOTA/graph.json","fetch_events":"https://pith.science/api/pith-number/EBKRLX26ILQ7L3KYP66RVVXOTA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA/action/storage_attestation","attest_author":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA/action/author_attestation","sign_citation":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA/action/citation_signature","submit_replication":"https://pith.science/pith/EBKRLX26ILQ7L3KYP66RVVXOTA/action/replication_record"}},"created_at":"2026-07-05T11:58:49.268955+00:00","updated_at":"2026-07-05T11:58:49.268955+00:00"}