{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QGCGJZ3FX5B3XBBFBQTYOWTZFR","short_pith_number":"pith:QGCGJZ3F","schema_version":"1.0","canonical_sha256":"818464e765bf43bb84250c27875a792c754c0bd1fd6e2fe3350400fb14148aec","source":{"kind":"arxiv","id":"2605.13817","version":1},"attestation_state":"computed","paper":{"title":"Neurosymbolic Auditing of Natural-Language Software Requirements","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Large language models with an SMT solver can translate and audit natural-language software requirements for ambiguity and defects.","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Bethel Hall, William Eiers","submitted_at":"2026-05-13T17:43:13Z","abstract_excerpt":"Natural-language software requirements are often ambiguous, inconsistent, and underspecified; in safety-critical domains, these defects propagate into formal models that verify the wrong specification and into implementations that ship unsafe behavior. We show that large language models, equipped with an SMT solver, can audit such requirements: translating them into formal logic, detecting ambiguity through stochastic variation in the generated formalization, and exposing inconsistency, vacuousness, and safety violations through solver queries on the resulting specification. We present VERIMED"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.13817","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-05-13T17:43:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f0d61d0272c1d967335b581877d07dc6caa75a70e94f9aaed34390edd5d12856","abstract_canon_sha256":"a1c4b7fe399eb389cce542ca3b1fbfa06caece94be51760c5b7bb82f8eb2e6c8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:15.313556Z","signature_b64":"T7oG/qUTGu8MRnKc2AjsXQUJAFSiuXTWoR8f2umxTp0iRRkR0n2IcPQVucGeqvIsIgIjppvZ8A4xJpPD80WzBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"818464e765bf43bb84250c27875a792c754c0bd1fd6e2fe3350400fb14148aec","last_reissued_at":"2026-05-18T02:44:15.312948Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:15.312948Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Neurosymbolic Auditing of Natural-Language Software Requirements","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Large language models with an SMT solver can translate and audit natural-language software requirements for ambiguity and defects.","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Bethel Hall, William Eiers","submitted_at":"2026-05-13T17:43:13Z","abstract_excerpt":"Natural-language software requirements are often ambiguous, inconsistent, and underspecified; in safety-critical domains, these defects propagate into formal models that verify the wrong specification and into implementations that ship unsafe behavior. We show that large language models, equipped with an SMT solver, can audit such requirements: translating them into formal logic, detecting ambiguity through stochastic variation in the generated formalization, and exposing inconsistency, vacuousness, and safety violations through solver queries on the resulting specification. We present VERIMED"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"large language models, equipped with an SMT solver, can audit such requirements: translating them into formal logic, detecting ambiguity through stochastic variation in the generated formalization, and exposing inconsistency, vacuousness, and safety violations through solver queries on the resulting specification.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"that stochastic variation in LLM-generated formalizations reliably signals genuine ambiguity in the original natural-language requirement rather than LLM inconsistency or hallucination.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"VERIMED translates natural-language requirements to formal logic via LLMs, detects ambiguity from stochastic formalization differences, and audits for inconsistency and safety violations using SMT queries.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Large language models with an SMT solver can translate and audit natural-language software requirements for ambiguity and defects.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"29cc8daff6d70b1ecbdcfd5565d71bb7b5cdada0ac7e6e36118e7b70ca3b2526"},"source":{"id":"2605.13817","kind":"arxiv","version":1},"verdict":{"id":"174b28b7-5524-49e6-98ea-c50276fb6284","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T17:57:24.470155Z","strongest_claim":"large language models, equipped with an SMT solver, can audit such requirements: translating them into formal logic, detecting ambiguity through stochastic variation in the generated formalization, and exposing inconsistency, vacuousness, and safety violations through solver queries on the resulting specification.","one_line_summary":"VERIMED translates natural-language requirements to formal logic via LLMs, detects ambiguity from stochastic formalization differences, and audits for inconsistency and safety violations using SMT queries.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"that stochastic variation in LLM-generated formalizations reliably signals genuine ambiguity in the original natural-language requirement rather than LLM inconsistency or hallucination.","pith_extraction_headline":"Large language models with an SMT solver can translate and audit natural-language software requirements for ambiguity and defects."},"references":{"count":29,"sample":[{"doi":"","year":2025,"title":"A. Akinfaderin and S. Subramanian. Verafi: Verified agentic financial intelligence through neurosymbolic policy generation.arXiv preprint arXiv:2512.14744, 2025. 9","work_id":"c9d4e896-4f48-464e-8a2a-b82be4055b7a","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"M. Allamanis, S. Panthaplackel, and P. Yin. Unsupervised evaluation of code llms with round-trip correctness.arXiv preprint arXiv:2402.08699, 2024","work_id":"7e0ec464-b6ec-48c2-9d2f-97bf9e2d1ff0","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Faithful Autoformalization via Roundtrip Verification and Repair","work_id":"f522d33d-3557-4b9e-ad69-f374edb17c3a","ref_index":3,"cited_arxiv_id":"2604.25031","is_internal_anchor":true},{"doi":"","year":2016,"title":"P. Arcaini, S. Bonfanti, A. Gargantini, and E. Riccobene. How to assure correctness and safety of medical software: the hemodialysis machine case study. InInternational Conference on Abstract State Ma","work_id":"807546e8-71ac-4b76-9c34-e7060dc516a2","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"C. Barrett, P. Fontaine, and C. Tinelli. The satisfiability modulo theories library (smt-lib). www. SMT-LIB. org, 2:68, 2016","work_id":"ba53710d-dbb0-4420-8e94-1e667da030df","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":29,"snapshot_sha256":"53db5cc6166def21d39bd5ab4e11c019d0148ff8c416c4bfa3dae5b1866e880a","internal_anchors":1},"formal_canon":{"evidence_count":1,"snapshot_sha256":"aaad35982e1b20ee4a257f848725c614a150efd1fe211abaeac2c414f0305799"},"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":"2605.13817","created_at":"2026-05-18T02:44:15.313038+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.13817v1","created_at":"2026-05-18T02:44:15.313038+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13817","created_at":"2026-05-18T02:44:15.313038+00:00"},{"alias_kind":"pith_short_12","alias_value":"QGCGJZ3FX5B3","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"QGCGJZ3FX5B3XBBF","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"QGCGJZ3F","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":1,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR","json":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR.json","graph_json":"https://pith.science/api/pith-number/QGCGJZ3FX5B3XBBFBQTYOWTZFR/graph.json","events_json":"https://pith.science/api/pith-number/QGCGJZ3FX5B3XBBFBQTYOWTZFR/events.json","paper":"https://pith.science/paper/QGCGJZ3F"},"agent_actions":{"view_html":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR","download_json":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR.json","view_paper":"https://pith.science/paper/QGCGJZ3F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.13817&json=true","fetch_graph":"https://pith.science/api/pith-number/QGCGJZ3FX5B3XBBFBQTYOWTZFR/graph.json","fetch_events":"https://pith.science/api/pith-number/QGCGJZ3FX5B3XBBFBQTYOWTZFR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR/action/storage_attestation","attest_author":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR/action/author_attestation","sign_citation":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR/action/citation_signature","submit_replication":"https://pith.science/pith/QGCGJZ3FX5B3XBBFBQTYOWTZFR/action/replication_record"}},"created_at":"2026-05-18T02:44:15.313038+00:00","updated_at":"2026-05-18T02:44:15.313038+00:00"}