{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:IY3ET7Y34L2B2CDZOXUMH2N4HW","short_pith_number":"pith:IY3ET7Y3","schema_version":"1.0","canonical_sha256":"463649ff1be2f41d087975e8c3e9bc3d8792aaa18834cf1451aacfa19cbe6fb9","source":{"kind":"arxiv","id":"2104.06678","version":1},"attestation_state":"computed","paper":{"title":"Large-Scale Self- and Semi-Supervised Learning for Speech Translation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexei Baevski, Alexis Conneau, Anne Wu, Changhan Wang, Juan Pino, Michael Auli","submitted_at":"2021-04-14T07:44:52Z","abstract_excerpt":"In this paper, we improve speech translation (ST) through effectively leveraging large quantities of unlabeled speech and text data in different and complementary ways. We explore both pretraining and self-training by using the large Libri-Light speech audio corpus and language modeling with CommonCrawl. Our experiments improve over the previous state of the art by 2.6 BLEU on average on all four considered CoVoST 2 language pairs via a simple recipe of combining wav2vec 2.0 pretraining, a single iteration of self-training and decoding with a language model. Different to existing work, our app"},"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":"2104.06678","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-04-14T07:44:52Z","cross_cats_sorted":[],"title_canon_sha256":"34f73e90077a6d2d99183a4fcdbbb0c6b0776389a2bad473169c6606d3e61dd9","abstract_canon_sha256":"70f7b09d7897b439aedcfbfb404d4b2b59f3500a43755d6fa00588120bf66230"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:32:08.340080Z","signature_b64":"yR8v7+E1EA8eJxr1QNq0kdu8+UlnMN0LHUkw+ll9p9SrRpD2euuUMR7OSZw5uq7QOeZ0JvO/Hnn6ubLdsmewBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"463649ff1be2f41d087975e8c3e9bc3d8792aaa18834cf1451aacfa19cbe6fb9","last_reissued_at":"2026-07-05T02:32:08.339701Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:32:08.339701Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Large-Scale Self- and Semi-Supervised Learning for Speech Translation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexei Baevski, Alexis Conneau, Anne Wu, Changhan Wang, Juan Pino, Michael Auli","submitted_at":"2021-04-14T07:44:52Z","abstract_excerpt":"In this paper, we improve speech translation (ST) through effectively leveraging large quantities of unlabeled speech and text data in different and complementary ways. We explore both pretraining and self-training by using the large Libri-Light speech audio corpus and language modeling with CommonCrawl. Our experiments improve over the previous state of the art by 2.6 BLEU on average on all four considered CoVoST 2 language pairs via a simple recipe of combining wav2vec 2.0 pretraining, a single iteration of self-training and decoding with a language model. Different to existing work, our app"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2104.06678","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/2104.06678/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":"2104.06678","created_at":"2026-07-05T02:32:08.339765+00:00"},{"alias_kind":"arxiv_version","alias_value":"2104.06678v1","created_at":"2026-07-05T02:32:08.339765+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2104.06678","created_at":"2026-07-05T02:32:08.339765+00:00"},{"alias_kind":"pith_short_12","alias_value":"IY3ET7Y34L2B","created_at":"2026-07-05T02:32:08.339765+00:00"},{"alias_kind":"pith_short_16","alias_value":"IY3ET7Y34L2B2CDZ","created_at":"2026-07-05T02:32:08.339765+00:00"},{"alias_kind":"pith_short_8","alias_value":"IY3ET7Y3","created_at":"2026-07-05T02:32:08.339765+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.00377","citing_title":"When End-to-End is Overkill: Rethinking Cascaded Speech-to-Text Translation","ref_index":6,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW","json":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW.json","graph_json":"https://pith.science/api/pith-number/IY3ET7Y34L2B2CDZOXUMH2N4HW/graph.json","events_json":"https://pith.science/api/pith-number/IY3ET7Y34L2B2CDZOXUMH2N4HW/events.json","paper":"https://pith.science/paper/IY3ET7Y3"},"agent_actions":{"view_html":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW","download_json":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW.json","view_paper":"https://pith.science/paper/IY3ET7Y3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2104.06678&json=true","fetch_graph":"https://pith.science/api/pith-number/IY3ET7Y34L2B2CDZOXUMH2N4HW/graph.json","fetch_events":"https://pith.science/api/pith-number/IY3ET7Y34L2B2CDZOXUMH2N4HW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW/action/storage_attestation","attest_author":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW/action/author_attestation","sign_citation":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW/action/citation_signature","submit_replication":"https://pith.science/pith/IY3ET7Y34L2B2CDZOXUMH2N4HW/action/replication_record"}},"created_at":"2026-07-05T02:32:08.339765+00:00","updated_at":"2026-07-05T02:32:08.339765+00:00"}