{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:F4NTWXVSIFA5R32GSI3OULVWQN","short_pith_number":"pith:F4NTWXVS","schema_version":"1.0","canonical_sha256":"2f1b3b5eb24141d8ef469236ea2eb68341e63b4eeda211d34e5a680df5353121","source":{"kind":"arxiv","id":"2305.11095","version":3},"attestation_state":"computed","paper":{"title":"Prompting the Hidden Talent of Web-Scale Speech Models for Zero-Shot Task Generalization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Brian Yan, David Harwath, Puyuan Peng, Shinji Watanabe","submitted_at":"2023-05-18T16:32:58Z","abstract_excerpt":"We investigate the emergent abilities of the recently proposed web-scale speech model Whisper, by adapting it to unseen tasks with prompt engineering. We selected three tasks: audio-visual speech recognition (AVSR), code-switched speech recognition (CS-ASR), and speech translation (ST) on unseen language pairs. We design task-specific prompts, by either leveraging another large-scale model, or simply manipulating the special tokens in the default prompts. Experiments show that compared to the default prompts, our proposed prompts improve performance by 10% to 45% on the three zero-shot tasks, "},"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":"2305.11095","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2023-05-18T16:32:58Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG","cs.SD"],"title_canon_sha256":"fa4bc11213f6a17666483dbb65f7dbcd739ab4bf1d751b0220576d0eeed74cdf","abstract_canon_sha256":"8c43a557a96f0bf5fe59bddcff33f915b04ea61ae0650f28db9c234713a99339"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:41:49.411292Z","signature_b64":"OdM/l3A2s8gFmo+UTNOaN/o8s/qEPrbPWYQWL/kcMSi8VQFNy3qHZd3bv85jLJGhryFsDwBgfR0+K237Jy3sDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2f1b3b5eb24141d8ef469236ea2eb68341e63b4eeda211d34e5a680df5353121","last_reissued_at":"2026-07-05T06:41:49.410806Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:41:49.410806Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Prompting the Hidden Talent of Web-Scale Speech Models for Zero-Shot Task Generalization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Brian Yan, David Harwath, Puyuan Peng, Shinji Watanabe","submitted_at":"2023-05-18T16:32:58Z","abstract_excerpt":"We investigate the emergent abilities of the recently proposed web-scale speech model Whisper, by adapting it to unseen tasks with prompt engineering. We selected three tasks: audio-visual speech recognition (AVSR), code-switched speech recognition (CS-ASR), and speech translation (ST) on unseen language pairs. We design task-specific prompts, by either leveraging another large-scale model, or simply manipulating the special tokens in the default prompts. Experiments show that compared to the default prompts, our proposed prompts improve performance by 10% to 45% on the three zero-shot tasks, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.11095","kind":"arxiv","version":3},"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/2305.11095/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":"2305.11095","created_at":"2026-07-05T06:41:49.410863+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.11095v3","created_at":"2026-07-05T06:41:49.410863+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.11095","created_at":"2026-07-05T06:41:49.410863+00:00"},{"alias_kind":"pith_short_12","alias_value":"F4NTWXVSIFA5","created_at":"2026-07-05T06:41:49.410863+00:00"},{"alias_kind":"pith_short_16","alias_value":"F4NTWXVSIFA5R32G","created_at":"2026-07-05T06:41:49.410863+00:00"},{"alias_kind":"pith_short_8","alias_value":"F4NTWXVS","created_at":"2026-07-05T06:41:49.410863+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.04491","citing_title":"Refining Transcripts With TV Subtitles by Prompt-Based Weakly Supervised Training of ASR","ref_index":13,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN","json":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN.json","graph_json":"https://pith.science/api/pith-number/F4NTWXVSIFA5R32GSI3OULVWQN/graph.json","events_json":"https://pith.science/api/pith-number/F4NTWXVSIFA5R32GSI3OULVWQN/events.json","paper":"https://pith.science/paper/F4NTWXVS"},"agent_actions":{"view_html":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN","download_json":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN.json","view_paper":"https://pith.science/paper/F4NTWXVS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.11095&json=true","fetch_graph":"https://pith.science/api/pith-number/F4NTWXVSIFA5R32GSI3OULVWQN/graph.json","fetch_events":"https://pith.science/api/pith-number/F4NTWXVSIFA5R32GSI3OULVWQN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN/action/storage_attestation","attest_author":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN/action/author_attestation","sign_citation":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN/action/citation_signature","submit_replication":"https://pith.science/pith/F4NTWXVSIFA5R32GSI3OULVWQN/action/replication_record"}},"created_at":"2026-07-05T06:41:49.410863+00:00","updated_at":"2026-07-05T06:41:49.410863+00:00"}