{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:HG5XWQAYQIHPEOBRXA5HCGO3UB","short_pith_number":"pith:HG5XWQAY","schema_version":"1.0","canonical_sha256":"39bb7b4018820ef23831b83a7119dba045255f4ff49c575714bcffc167dbc142","source":{"kind":"arxiv","id":"2502.12366","version":1},"attestation_state":"computed","paper":{"title":"ScriptoriumWS: A Code Generation Assistant for Weak Supervision","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Catherine Cao, Frederic Sala, Harit Vishwakarma, Nicholas Roberts, Spencer Schoenberg, Tzu-Heng Huang","submitted_at":"2025-02-17T23:07:14Z","abstract_excerpt":"Weak supervision is a popular framework for overcoming the labeled data bottleneck: the need to obtain labels for training data. In weak supervision, multiple noisy-but-cheap sources are used to provide guesses of the label and are aggregated to produce high-quality pseudolabels. These sources are often expressed as small programs written by domain experts -- and so are expensive to obtain. Instead, we argue for using code-generation models to act as coding assistants for crafting weak supervision sources. We study prompting strategies to maximize the quality of the generated sources, settling"},"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":"2502.12366","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-17T23:07:14Z","cross_cats_sorted":[],"title_canon_sha256":"191c89235e8296e2787fa28e01e162420f88360a1762deb61e5765d9b92a0708","abstract_canon_sha256":"95a0e2b69869547de2f76675ce81992056f130a7c411cd5a3a2f78dfee7f0a67"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:15:54.511760Z","signature_b64":"ZlNLLTX/6s4WquM1L4e3+u4PwOPtTn+u5iEbLmeHmVKKpEbQwCd197gUw9cyaXDFIIRpqQRw+sd9Zaml6EcoBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"39bb7b4018820ef23831b83a7119dba045255f4ff49c575714bcffc167dbc142","last_reissued_at":"2026-07-05T10:15:54.511263Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:15:54.511263Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ScriptoriumWS: A Code Generation Assistant for Weak Supervision","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Catherine Cao, Frederic Sala, Harit Vishwakarma, Nicholas Roberts, Spencer Schoenberg, Tzu-Heng Huang","submitted_at":"2025-02-17T23:07:14Z","abstract_excerpt":"Weak supervision is a popular framework for overcoming the labeled data bottleneck: the need to obtain labels for training data. In weak supervision, multiple noisy-but-cheap sources are used to provide guesses of the label and are aggregated to produce high-quality pseudolabels. These sources are often expressed as small programs written by domain experts -- and so are expensive to obtain. Instead, we argue for using code-generation models to act as coding assistants for crafting weak supervision sources. We study prompting strategies to maximize the quality of the generated sources, settling"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.12366","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/2502.12366/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":"2502.12366","created_at":"2026-07-05T10:15:54.511325+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.12366v1","created_at":"2026-07-05T10:15:54.511325+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.12366","created_at":"2026-07-05T10:15:54.511325+00:00"},{"alias_kind":"pith_short_12","alias_value":"HG5XWQAYQIHP","created_at":"2026-07-05T10:15:54.511325+00:00"},{"alias_kind":"pith_short_16","alias_value":"HG5XWQAYQIHPEOBR","created_at":"2026-07-05T10:15:54.511325+00:00"},{"alias_kind":"pith_short_8","alias_value":"HG5XWQAY","created_at":"2026-07-05T10:15:54.511325+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.10403","citing_title":"Time To Impeach LLM-as-a-Judge: Programs are the Future of Evaluation","ref_index":32,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB","json":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB.json","graph_json":"https://pith.science/api/pith-number/HG5XWQAYQIHPEOBRXA5HCGO3UB/graph.json","events_json":"https://pith.science/api/pith-number/HG5XWQAYQIHPEOBRXA5HCGO3UB/events.json","paper":"https://pith.science/paper/HG5XWQAY"},"agent_actions":{"view_html":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB","download_json":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB.json","view_paper":"https://pith.science/paper/HG5XWQAY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.12366&json=true","fetch_graph":"https://pith.science/api/pith-number/HG5XWQAYQIHPEOBRXA5HCGO3UB/graph.json","fetch_events":"https://pith.science/api/pith-number/HG5XWQAYQIHPEOBRXA5HCGO3UB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB/action/storage_attestation","attest_author":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB/action/author_attestation","sign_citation":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB/action/citation_signature","submit_replication":"https://pith.science/pith/HG5XWQAYQIHPEOBRXA5HCGO3UB/action/replication_record"}},"created_at":"2026-07-05T10:15:54.511325+00:00","updated_at":"2026-07-05T10:15:54.511325+00:00"}