{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:CHNUWEDWDP63SFWWLT2734TDLY","short_pith_number":"pith:CHNUWEDW","schema_version":"1.0","canonical_sha256":"11db4b10761bfdb916d65cf5fdf2635e3704b0b3c4b270e28c3009ed2033c7fe","source":{"kind":"arxiv","id":"2205.12446","version":1},"attestation_state":"computed","paper":{"title":"FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Alexis Conneau, Ankur Bapna, Clara Rivera, Jason Riesa, Min Ma, Siddharth Dalmia, Simran Khanuja, Vera Axelrod, Yu Zhang","submitted_at":"2022-05-25T02:29:03Z","abstract_excerpt":"We introduce FLEURS, the Few-shot Learning Evaluation of Universal Representations of Speech benchmark. FLEURS is an n-way parallel speech dataset in 102 languages built on top of the machine translation FLoRes-101 benchmark, with approximately 12 hours of speech supervision per language. FLEURS can be used for a variety of speech tasks, including Automatic Speech Recognition (ASR), Speech Language Identification (Speech LangID), Translation and Retrieval. In this paper, we provide baselines for the tasks based on multilingual pre-trained models like mSLAM. The goal of FLEURS is to enable spee"},"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":"2205.12446","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-05-25T02:29:03Z","cross_cats_sorted":["cs.LG","cs.SD","eess.AS"],"title_canon_sha256":"25a37c297c3b2e27efebc4e78d84b48f5d1cdf1541b91c389f3c4abffbf5a85d","abstract_canon_sha256":"221b16bd34b9a6b278b0eb958f4efb7cf4941839525b6e9d0b8a7af26231c811"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:26:25.787675Z","signature_b64":"++QjTgPh5l2RHd778vnGpPGV3MsZIXzoEFDrvtW4EyzWRltyQo5oHkfHg2KC/NRb9zVzqtqD5+4AsKncDWGdCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"11db4b10761bfdb916d65cf5fdf2635e3704b0b3c4b270e28c3009ed2033c7fe","last_reissued_at":"2026-07-05T04:26:25.787252Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:26:25.787252Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.SD","eess.AS"],"primary_cat":"cs.CL","authors_text":"Alexis Conneau, Ankur Bapna, Clara Rivera, Jason Riesa, Min Ma, Siddharth Dalmia, Simran Khanuja, Vera Axelrod, Yu Zhang","submitted_at":"2022-05-25T02:29:03Z","abstract_excerpt":"We introduce FLEURS, the Few-shot Learning Evaluation of Universal Representations of Speech benchmark. FLEURS is an n-way parallel speech dataset in 102 languages built on top of the machine translation FLoRes-101 benchmark, with approximately 12 hours of speech supervision per language. FLEURS can be used for a variety of speech tasks, including Automatic Speech Recognition (ASR), Speech Language Identification (Speech LangID), Translation and Retrieval. In this paper, we provide baselines for the tasks based on multilingual pre-trained models like mSLAM. The goal of FLEURS is to enable spee"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.12446","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/2205.12446/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":"2205.12446","created_at":"2026-07-05T04:26:25.787307+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.12446v1","created_at":"2026-07-05T04:26:25.787307+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.12446","created_at":"2026-07-05T04:26:25.787307+00:00"},{"alias_kind":"pith_short_12","alias_value":"CHNUWEDWDP63","created_at":"2026-07-05T04:26:25.787307+00:00"},{"alias_kind":"pith_short_16","alias_value":"CHNUWEDWDP63SFWW","created_at":"2026-07-05T04:26:25.787307+00:00"},{"alias_kind":"pith_short_8","alias_value":"CHNUWEDW","created_at":"2026-07-05T04:26:25.787307+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":13,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.06827","citing_title":"Compress the Cache, Not the Speech Embedding: KV Compression for Efficient Speech LLMs","ref_index":49,"is_internal_anchor":true},{"citing_arxiv_id":"2606.09335","citing_title":"Factors affecting ASR performance: A study using state of the art ASR models in Indic Languages","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13087","citing_title":"Vividh-ASR: A Complexity-Tiered Benchmark and Optimization Dynamics for Robust Indic Speech Recognition","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25596","citing_title":"Multilingual Phonological Feature Recognition with Self-Supervised Speech Models","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28211","citing_title":"When Helpful Context Leaks: Privacy Risks in Domain-Adapted ASR","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23463","citing_title":"StepAudio 2.5 Technical Report","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2512.01512","citing_title":"MCAT: Scaling Many-to-Many Speech-to-Text Translation with MLLMs to 70 Languages","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2512.16378","citing_title":"Hearing to Translate: The Effectiveness of Speech Modality Integration into LLMs","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13087","citing_title":"Vividh-ASR: A Complexity-Tiered Benchmark and Optimization Dynamics for Robust Indic Speech Recognition","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19565","citing_title":"Detecting Hallucinations in SpeechLLMs at Inference Time Using Attention Maps","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19221","citing_title":"UAF: A Unified Audio Front-end LLM for Full-Duplex Speech Interaction","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10736","citing_title":"BlasBench: An Open Benchmark for Irish Speech Recognition","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04598","citing_title":"Benchmarking Multilingual Speech Models on Pashto: Zero-Shot ASR, Script Failure, and Cross-Domain Evaluation","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY","json":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY.json","graph_json":"https://pith.science/api/pith-number/CHNUWEDWDP63SFWWLT2734TDLY/graph.json","events_json":"https://pith.science/api/pith-number/CHNUWEDWDP63SFWWLT2734TDLY/events.json","paper":"https://pith.science/paper/CHNUWEDW"},"agent_actions":{"view_html":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY","download_json":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY.json","view_paper":"https://pith.science/paper/CHNUWEDW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.12446&json=true","fetch_graph":"https://pith.science/api/pith-number/CHNUWEDWDP63SFWWLT2734TDLY/graph.json","fetch_events":"https://pith.science/api/pith-number/CHNUWEDWDP63SFWWLT2734TDLY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY/action/storage_attestation","attest_author":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY/action/author_attestation","sign_citation":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY/action/citation_signature","submit_replication":"https://pith.science/pith/CHNUWEDWDP63SFWWLT2734TDLY/action/replication_record"}},"created_at":"2026-07-05T04:26:25.787307+00:00","updated_at":"2026-07-05T04:26:25.787307+00:00"}