{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:JMGKQMSGLCJKFGTIOLIWXCNN5H","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"df34a9e994c397bc27c0cf30ea9fee8bd38fec945293dcfd864559e1e12ccc7d","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.GR","submitted_at":"2026-05-17T13:47:38Z","title_canon_sha256":"69a58509d33aa03f38302fda1996a48ca665746c4b7b9714778dbedb72324e1a"},"schema_version":"1.0","source":{"id":"2605.17448","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.17448","created_at":"2026-05-20T00:04:39Z"},{"alias_kind":"arxiv_version","alias_value":"2605.17448v1","created_at":"2026-05-20T00:04:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17448","created_at":"2026-05-20T00:04:39Z"},{"alias_kind":"pith_short_12","alias_value":"JMGKQMSGLCJK","created_at":"2026-05-20T00:04:39Z"},{"alias_kind":"pith_short_16","alias_value":"JMGKQMSGLCJKFGTI","created_at":"2026-05-20T00:04:39Z"},{"alias_kind":"pith_short_8","alias_value":"JMGKQMSG","created_at":"2026-05-20T00:04:39Z"}],"graph_snapshots":[{"event_id":"sha256:f028a273cf29b122e99f2c5bd3592ca13a16a266d1789c11188b081c9561e303","target":"graph","created_at":"2026-05-20T00:04:39Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Codex (GPT-5.5) and Claude Code (Opus-4.7) agents do not produce a single strict-passing artifact in the main first-attempt sweep, with the best configuration meeting only about 20% of typed requirements on average. The same feedback tools improve geometric reconstruction, with GPT-5.5/xhigh rising from 0.444 to 0.592 Box-IoU on S2O and from 0.397 to 0.505 on Fusion360."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That finite element analysis on the generated STEP files provides a reliable and sufficient proxy for real engineering requirements and that the observed IoU gains are causally due to the new blueprint and image feedback rather than other unmeasured factors in the agent loop. This premise enters in the abstract's description of the validation process and the reported improvements."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"CAD agents using finite element analysis feedback plus new text blueprint and multi-view image signals improve geometric accuracy on S2O and Fusion360 benchmarks while addressing physical validity gaps in prior generation methods."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"CAD agents improve designs when finite element analysis and blueprint feedback close the loop between generation and engineering checks."}],"snapshot_sha256":"c2e0a214561114797110e7152383f98a64f07c87b17bf1461410556d14093252"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"4db5869901352b483aafe88b0571a393e968501b8973c4bc83fe17a7ac2fa454"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T23:01:19.590100Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T22:51:52.265509Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T21:41:57.716191Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T21:33:23.668069Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.17448/integrity.json","findings":[],"snapshot_sha256":"dea68ae54f62a1d5775799bee4dfb762a10260892affd021fbd339c104357e42","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Computer-aided design (CAD) is the backbone of modern industrial design, yet learned CAD generators still fall short of real engineering pipelines: they neither iterate like engineers nor evaluate what engineering requires. Prior work has treated CAD generation as two disjoint steps, part synthesis and assembly, where the former is graded by proximity to a gold reference and the latter, when handled at all, is reduced to a separate constraint solving step. In this work, we introduce a more industry-native task formulation that requires a model to produce a fully assembled multi-part STEP file ","authors_text":"Guijin Son, Jehyun Park, Seyeon Park, Sunghee Ahn, Youngjae Yu","cross_cats":["cs.CL"],"headline":"CAD agents improve designs when finite element analysis and blueprint feedback close the loop between generation and engineering checks.","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.GR","submitted_at":"2026-05-17T13:47:38Z","title":"Self-Improving CAD Generation Agents with Finite Element Analysis as Feedback"},"references":{"count":37,"internal_anchors":3,"resolved_work":37,"sample":[{"cited_arxiv_id":"2508.10925","doi":"","is_internal_anchor":true,"ref_index":1,"title":"gpt-oss-120b & gpt-oss-20b Model Card","work_id":"178c1f7e-4f19-4392-a45d-45a6dfa88ead","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Generating cad code with vision-language models for 3d designs","work_id":"8530360e-db93-415e-a799-6bf6cdfc7690","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Developing a computer use model","work_id":"f5ee6ef4-18e8-42de-a934-9deca47f323d","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Cadsmith: Multi-agent cad genera- tion with programmatic geometric validation.arXiv preprint arXiv:2603.26512, 2026","work_id":"6678d33e-c232-4904-aabd-8bdc2010f19a","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Geoffrey Boothroyd, Peter Dewhurst, and Winston A Knight.Product design for manufacture and assembly. CRC press, 2010","work_id":"2862922d-c75e-426a-9334-f80f549a518d","year":2010}],"snapshot_sha256":"84916e83e20221fcf34074ce7f1b66db219ff8bf1c45fd2969717457b46a0781"},"source":{"id":"2605.17448","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T22:36:46.130529Z","id":"2b5de34e-1d9a-41e2-9323-126a449b8df3","model_set":{"reader":"grok-4.3"},"one_line_summary":"CAD agents using finite element analysis feedback plus new text blueprint and multi-view image signals improve geometric accuracy on S2O and Fusion360 benchmarks while addressing physical validity gaps in prior generation methods.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"CAD agents improve designs when finite element analysis and blueprint feedback close the loop between generation and engineering checks.","strongest_claim":"Codex (GPT-5.5) and Claude Code (Opus-4.7) agents do not produce a single strict-passing artifact in the main first-attempt sweep, with the best configuration meeting only about 20% of typed requirements on average. The same feedback tools improve geometric reconstruction, with GPT-5.5/xhigh rising from 0.444 to 0.592 Box-IoU on S2O and from 0.397 to 0.505 on Fusion360.","weakest_assumption":"That finite element analysis on the generated STEP files provides a reliable and sufficient proxy for real engineering requirements and that the observed IoU gains are causally due to the new blueprint and image feedback rather than other unmeasured factors in the agent loop. This premise enters in the abstract's description of the validation process and the reported improvements."}},"verdict_id":"2b5de34e-1d9a-41e2-9323-126a449b8df3"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:ed67907e4dac8836a7a39e94c7a8ec169fc5b0a8b37bfc8950326e6fef9547df","target":"record","created_at":"2026-05-20T00:04:39Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"df34a9e994c397bc27c0cf30ea9fee8bd38fec945293dcfd864559e1e12ccc7d","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.GR","submitted_at":"2026-05-17T13:47:38Z","title_canon_sha256":"69a58509d33aa03f38302fda1996a48ca665746c4b7b9714778dbedb72324e1a"},"schema_version":"1.0","source":{"id":"2605.17448","kind":"arxiv","version":1}},"canonical_sha256":"4b0ca832465892a29a6872d16b89ade9f8d963409e66c46659361a197872bcaf","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4b0ca832465892a29a6872d16b89ade9f8d963409e66c46659361a197872bcaf","first_computed_at":"2026-05-20T00:04:39.500379Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:04:39.500379Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MgmYfRx/pomSiZBoHG655WQgNcrFXXuqmRNllRmSa84j0BzMl6UrT5gA4uoRvPZrAIxiyLmtRDiAGybkPz9nCw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:04:39.501154Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.17448","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ed67907e4dac8836a7a39e94c7a8ec169fc5b0a8b37bfc8950326e6fef9547df","sha256:f028a273cf29b122e99f2c5bd3592ca13a16a266d1789c11188b081c9561e303"],"state_sha256":"1db2c9fbb3d7bd425611c586a5d689c4066c1c61afdf558a2775479e8f173792"}