{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:KWGSTDY4YZUISQDYFO3RQQSALM","short_pith_number":"pith:KWGSTDY4","schema_version":"1.0","canonical_sha256":"558d298f1cc6688940782bb71842405b2851b771fc669f999866dcd5534c9ab0","source":{"kind":"arxiv","id":"2305.15663","version":1},"attestation_state":"computed","paper":{"title":"Mixture-of-Expert Conformer for Streaming Multilingual ASR","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Bo Li, Francoise Beaufays, Ke Hu, Tara N. Sainath, Yu Zhang","submitted_at":"2023-05-25T02:16:32Z","abstract_excerpt":"End-to-end models with large capacity have significantly improved multilingual automatic speech recognition, but their computation cost poses challenges for on-device applications. We propose a streaming truly multilingual Conformer incorporating mixture-of-expert (MoE) layers that learn to only activate a subset of parameters in training and inference. The MoE layer consists of a softmax gate which chooses the best two experts among many in forward propagation. The proposed MoE layer offers efficient inference by activating a fixed number of parameters as the number of experts increases. We e"},"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":"2305.15663","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-25T02:16:32Z","cross_cats_sorted":["cs.SD","eess.AS"],"title_canon_sha256":"66c02806d52f70b47f8b65dbc37083c9d0e339da97c88404eafec26259f25920","abstract_canon_sha256":"cca8aa00be4a1cb65b13130a55da0307c76a05c69902fd9a2ac63d7a82adda60"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:13:42.967266Z","signature_b64":"2I8bLTgyqOVVqR38lpqZm1ElHb5SUAKC8F5tNoBBO1ykL9OGt91XWXBmxmTVcFlrYW9r2JnmI2/Zk+W6tUbTDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"558d298f1cc6688940782bb71842405b2851b771fc669f999866dcd5534c9ab0","last_reissued_at":"2026-07-05T06:13:42.966820Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:13:42.966820Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mixture-of-Expert Conformer for Streaming Multilingual ASR","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Bo Li, Francoise Beaufays, Ke Hu, Tara N. Sainath, Yu Zhang","submitted_at":"2023-05-25T02:16:32Z","abstract_excerpt":"End-to-end models with large capacity have significantly improved multilingual automatic speech recognition, but their computation cost poses challenges for on-device applications. We propose a streaming truly multilingual Conformer incorporating mixture-of-expert (MoE) layers that learn to only activate a subset of parameters in training and inference. The MoE layer consists of a softmax gate which chooses the best two experts among many in forward propagation. The proposed MoE layer offers efficient inference by activating a fixed number of parameters as the number of experts increases. We e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.15663","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/2305.15663/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":"2305.15663","created_at":"2026-07-05T06:13:42.966884+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.15663v1","created_at":"2026-07-05T06:13:42.966884+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.15663","created_at":"2026-07-05T06:13:42.966884+00:00"},{"alias_kind":"pith_short_12","alias_value":"KWGSTDY4YZUI","created_at":"2026-07-05T06:13:42.966884+00:00"},{"alias_kind":"pith_short_16","alias_value":"KWGSTDY4YZUISQDY","created_at":"2026-07-05T06:13:42.966884+00:00"},{"alias_kind":"pith_short_8","alias_value":"KWGSTDY4","created_at":"2026-07-05T06:13:42.966884+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.15622","citing_title":"TouchASP: Elastic Automatic Speech Perception that Everyone Can Touch","ref_index":8,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM","json":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM.json","graph_json":"https://pith.science/api/pith-number/KWGSTDY4YZUISQDYFO3RQQSALM/graph.json","events_json":"https://pith.science/api/pith-number/KWGSTDY4YZUISQDYFO3RQQSALM/events.json","paper":"https://pith.science/paper/KWGSTDY4"},"agent_actions":{"view_html":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM","download_json":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM.json","view_paper":"https://pith.science/paper/KWGSTDY4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.15663&json=true","fetch_graph":"https://pith.science/api/pith-number/KWGSTDY4YZUISQDYFO3RQQSALM/graph.json","fetch_events":"https://pith.science/api/pith-number/KWGSTDY4YZUISQDYFO3RQQSALM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM/action/storage_attestation","attest_author":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM/action/author_attestation","sign_citation":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM/action/citation_signature","submit_replication":"https://pith.science/pith/KWGSTDY4YZUISQDYFO3RQQSALM/action/replication_record"}},"created_at":"2026-07-05T06:13:42.966884+00:00","updated_at":"2026-07-05T06:13:42.966884+00:00"}