{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:SUXQ7RETEU4DAWVRQ7MAFVSKJS","short_pith_number":"pith:SUXQ7RET","schema_version":"1.0","canonical_sha256":"952f0fc4932538305ab187d802d64a4cbeb85f65af4f8e8820d973d95d797ed7","source":{"kind":"arxiv","id":"2202.01374","version":1},"attestation_state":"computed","paper":{"title":"mSLAM: Massively multilingual joint pre-training for speech and text","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Alexis Conneau, Ankur Bapna, Colin Cherry, Jason Riesa, Melvin Johnson, Simran Khanuja, Ye Jia, Yong Cheng, Yu Zhang","submitted_at":"2022-02-03T02:26:40Z","abstract_excerpt":"We present mSLAM, a multilingual Speech and LAnguage Model that learns cross-lingual cross-modal representations of speech and text by pre-training jointly on large amounts of unlabeled speech and text in multiple languages. mSLAM combines w2v-BERT pre-training on speech with SpanBERT pre-training on character-level text, along with Connectionist Temporal Classification (CTC) losses on paired speech and transcript data, to learn a single model capable of learning from and representing both speech and text signals in a shared representation space. We evaluate mSLAM on several downstream speech "},"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":"2202.01374","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-02-03T02:26:40Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4e225b012c4fedc7428c94f59b730030b533f757caf95bc2d39ee0d1c5917f27","abstract_canon_sha256":"5c3f86b6ad9436156740ea0d4b4f6e6227cec696765f1cf2ddf78ff6d2e58ca5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:53:53.598607Z","signature_b64":"O+AHoKzpVUESZhXENTm63PxWyCQSjsMWJHfBmJwafqL85HOh6tEHcnr29AfvmdB71YMVTmH47X/wDgwEcyvhDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"952f0fc4932538305ab187d802d64a4cbeb85f65af4f8e8820d973d95d797ed7","last_reissued_at":"2026-07-05T03:53:53.598142Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:53:53.598142Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"mSLAM: Massively multilingual joint pre-training for speech and text","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Alexis Conneau, Ankur Bapna, Colin Cherry, Jason Riesa, Melvin Johnson, Simran Khanuja, Ye Jia, Yong Cheng, Yu Zhang","submitted_at":"2022-02-03T02:26:40Z","abstract_excerpt":"We present mSLAM, a multilingual Speech and LAnguage Model that learns cross-lingual cross-modal representations of speech and text by pre-training jointly on large amounts of unlabeled speech and text in multiple languages. mSLAM combines w2v-BERT pre-training on speech with SpanBERT pre-training on character-level text, along with Connectionist Temporal Classification (CTC) losses on paired speech and transcript data, to learn a single model capable of learning from and representing both speech and text signals in a shared representation space. We evaluate mSLAM on several downstream speech "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.01374","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/2202.01374/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":"2202.01374","created_at":"2026-07-05T03:53:53.598202+00:00"},{"alias_kind":"arxiv_version","alias_value":"2202.01374v1","created_at":"2026-07-05T03:53:53.598202+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.01374","created_at":"2026-07-05T03:53:53.598202+00:00"},{"alias_kind":"pith_short_12","alias_value":"SUXQ7RETEU4D","created_at":"2026-07-05T03:53:53.598202+00:00"},{"alias_kind":"pith_short_16","alias_value":"SUXQ7RETEU4DAWVR","created_at":"2026-07-05T03:53:53.598202+00:00"},{"alias_kind":"pith_short_8","alias_value":"SUXQ7RET","created_at":"2026-07-05T03:53:53.598202+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2512.16378","citing_title":"Hearing to Translate: The Effectiveness of Speech Modality Integration into LLMs","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2204.00598","citing_title":"Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2306.12925","citing_title":"AudioPaLM: A Large Language Model That Can Speak and Listen","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.26514","citing_title":"Text-Utilization for Encoder-dominated Speech Recognition Models","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS","json":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS.json","graph_json":"https://pith.science/api/pith-number/SUXQ7RETEU4DAWVRQ7MAFVSKJS/graph.json","events_json":"https://pith.science/api/pith-number/SUXQ7RETEU4DAWVRQ7MAFVSKJS/events.json","paper":"https://pith.science/paper/SUXQ7RET"},"agent_actions":{"view_html":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS","download_json":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS.json","view_paper":"https://pith.science/paper/SUXQ7RET","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2202.01374&json=true","fetch_graph":"https://pith.science/api/pith-number/SUXQ7RETEU4DAWVRQ7MAFVSKJS/graph.json","fetch_events":"https://pith.science/api/pith-number/SUXQ7RETEU4DAWVRQ7MAFVSKJS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS/action/storage_attestation","attest_author":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS/action/author_attestation","sign_citation":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS/action/citation_signature","submit_replication":"https://pith.science/pith/SUXQ7RETEU4DAWVRQ7MAFVSKJS/action/replication_record"}},"created_at":"2026-07-05T03:53:53.598202+00:00","updated_at":"2026-07-05T03:53:53.598202+00:00"}