{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:PMXUX6XS4QUPXI5ZL3ER56K4ET","short_pith_number":"pith:PMXUX6XS","schema_version":"1.0","canonical_sha256":"7b2f4bfaf2e428fba3b95ec91ef95c24e6321e7a50759e26c2cc3b6aa9d0201a","source":{"kind":"arxiv","id":"2604.00660","version":2},"attestation_state":"computed","paper":{"title":"Streaming Model Cascades for Semantic SQL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.DB","authors_text":"Kyle Schmaus, Pawe{\\l} Liskowski","submitted_at":"2026-04-01T09:07:08Z","abstract_excerpt":"Modern data warehouses extend SQL with semantic operators that invoke large language models on each qualifying row, making per-row inference orders of magnitude more expensive than traditional SQL. Model cascades reduce this cost by routing most rows through a fast proxy model and delegating uncertain cases to an expensive oracle. Prior SUPG-style cascades, however, require a global proxy-score pass that is itself an LLM-inference workload and blocks output in pipelined query engines. They also target either precision or recall and cannot serve workloads that need both. We formalize the cascad"},"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":"2604.00660","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DB","submitted_at":"2026-04-01T09:07:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9e1275f7b7c7feb3658bc53860aa4df7c34bf4844ae791db5d7d61010bf8b2f4","abstract_canon_sha256":"a5fd2f3dc597040c4fdabdc0050a0265b025c36ab7103c9be58ddd7b30a924a6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T01:16:11.237378Z","signature_b64":"81i9H0zRg5n6dnOCWMJ5d/BbHf1km26oe36hqNpi5KS3w/A27y3tTnuRYWLdLCVieuHZEOa4skyaNFYdCpw3CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b2f4bfaf2e428fba3b95ec91ef95c24e6321e7a50759e26c2cc3b6aa9d0201a","last_reissued_at":"2026-07-07T01:16:11.236481Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T01:16:11.236481Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Streaming Model Cascades for Semantic SQL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.DB","authors_text":"Kyle Schmaus, Pawe{\\l} Liskowski","submitted_at":"2026-04-01T09:07:08Z","abstract_excerpt":"Modern data warehouses extend SQL with semantic operators that invoke large language models on each qualifying row, making per-row inference orders of magnitude more expensive than traditional SQL. Model cascades reduce this cost by routing most rows through a fast proxy model and delegating uncertain cases to an expensive oracle. Prior SUPG-style cascades, however, require a global proxy-score pass that is itself an LLM-inference workload and blocks output in pipelined query engines. They also target either precision or recall and cannot serve workloads that need both. We formalize the cascad"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.00660","kind":"arxiv","version":2},"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/2604.00660/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":"2604.00660","created_at":"2026-07-07T01:16:11.236582+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.00660v2","created_at":"2026-07-07T01:16:11.236582+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.00660","created_at":"2026-07-07T01:16:11.236582+00:00"},{"alias_kind":"pith_short_12","alias_value":"PMXUX6XS4QUP","created_at":"2026-07-07T01:16:11.236582+00:00"},{"alias_kind":"pith_short_16","alias_value":"PMXUX6XS4QUPXI5Z","created_at":"2026-07-07T01:16:11.236582+00:00"},{"alias_kind":"pith_short_8","alias_value":"PMXUX6XS","created_at":"2026-07-07T01:16:11.236582+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2606.07923","citing_title":"Larch: Learned Query Optimization for Semantic Predicates","ref_index":35,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET","json":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET.json","graph_json":"https://pith.science/api/pith-number/PMXUX6XS4QUPXI5ZL3ER56K4ET/graph.json","events_json":"https://pith.science/api/pith-number/PMXUX6XS4QUPXI5ZL3ER56K4ET/events.json","paper":"https://pith.science/paper/PMXUX6XS"},"agent_actions":{"view_html":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET","download_json":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET.json","view_paper":"https://pith.science/paper/PMXUX6XS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.00660&json=true","fetch_graph":"https://pith.science/api/pith-number/PMXUX6XS4QUPXI5ZL3ER56K4ET/graph.json","fetch_events":"https://pith.science/api/pith-number/PMXUX6XS4QUPXI5ZL3ER56K4ET/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET/action/storage_attestation","attest_author":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET/action/author_attestation","sign_citation":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET/action/citation_signature","submit_replication":"https://pith.science/pith/PMXUX6XS4QUPXI5ZL3ER56K4ET/action/replication_record"}},"created_at":"2026-07-07T01:16:11.236582+00:00","updated_at":"2026-07-07T01:16:11.236582+00:00"}