In this work, we introduce a more industry-native task formulation that requires a model to produce a fully assembled multi-part STEP file from a free-form engineering brief, which is then validated via finite element analysis (FEA).<br><a href=\"https://cdn-uploads.huggingface.co/production/uploads/66cd506100cddde87a683668/1Y7qf81zRErYSGOyfVpaR.jpeg\" rel=\"nofollow\"><img src=\"https://cdn-uploads.huggingface.co/production/uploads/66cd506100cddde87a683668/1Y7qf81zRErYSGOyfVpaR.jpeg\" alt=\"IMG_9793\"></a></p>\n","updatedAt":"2026-05-26T00:41:46.028Z","author":{"_id":"66cd506100cddde87a683668","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/b1GWC2hU7ommj51pGZmu_.jpeg","fullname":"Jehyun Park","name":"Parkprogrammer","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":1,"identifiedLanguage":{"language":"en","probability":0.7502140998840332},"editors":["Parkprogrammer"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/b1GWC2hU7ommj51pGZmu_.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2605.17448","authors":[{"_id":"6a13f0e54d9e8d8602d20364","name":"Guijin Son","hidden":false},{"_id":"6a13f0e54d9e8d8602d20365","user":{"_id":"66cd506100cddde87a683668","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/b1GWC2hU7ommj51pGZmu_.jpeg","isPro":false,"fullname":"Jehyun Park","user":"Parkprogrammer","type":"user","name":"Parkprogrammer"},"name":"Jehyun Park","status":"claimed_verified","statusLastChangedAt":"2026-05-25T15:11:42.359Z","hidden":false},{"_id":"6a13f0e54d9e8d8602d20366","name":"Seyeon Park","hidden":false},{"_id":"6a13f0e54d9e8d8602d20367","name":"Sunghee Ahn","hidden":false},{"_id":"6a13f0e54d9e8d8602d20368","name":"Youngjae Yu","hidden":false}],"publishedAt":"2026-05-17T00:00:00.000Z","submittedOnDailyAt":"2026-05-25T00:00:00.000Z","title":"Self-Improving CAD Generation Agents with Finite Element Analysis as Feedback","submittedOnDailyBy":{"_id":"66cd506100cddde87a683668","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/b1GWC2hU7ommj51pGZmu_.jpeg","isPro":false,"fullname":"Jehyun Park","user":"Parkprogrammer","type":"user","name":"Parkprogrammer"},"summary":"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 from a free-form engineering brief, which is then validated via finite element analysis (FEA). FEA validation reveals that 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. Moreover, we introduce two additional supervision signals, a novel text-only blueprint schema and a 21-view image renderer that aids the agent's visual inspection, that better align the generation loop with how engineers iterate in practice. On S2O and Fusion360, 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. Together these signals move CAD programs toward artifacts that are not only visually plausible but also checked against physical and structural requirements.","upvotes":1,"discussionId":"6a13f0e54d9e8d8602d20369","ai_summary":"Learning-based CAD generation systems are enhanced by incorporating engineering validation through finite element analysis and improved supervision signals that better align with actual design processes.","ai_keywords":["computer-aided design","CAD generation","finite element analysis","STEP file","geometric reconstruction","Box-IoU"],"organization":{"_id":"66d54dc8033492801db2bf5a","name":"SeoulNatlUniv","fullname":"Seoul National University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/659ccc9d18897eb6594e897f/_-0BM-1UyM-d-lRiahFnf.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"66cd506100cddde87a683668","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/b1GWC2hU7ommj51pGZmu_.jpeg","isPro":false,"fullname":"Jehyun Park","user":"Parkprogrammer","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"66d54dc8033492801db2bf5a","name":"SeoulNatlUniv","fullname":"Seoul National University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/659ccc9d18897eb6594e897f/_-0BM-1UyM-d-lRiahFnf.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2605/2605.17448.md"}">
Self-Improving CAD Generation Agents with Finite Element Analysis as Feedback
Abstract
Learning-based CAD generation systems are enhanced by incorporating engineering validation through finite element analysis and improved supervision signals that better align with actual design processes.
AI-generated summary
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 from a free-form engineering brief, which is then validated via finite element analysis (FEA). FEA validation reveals that 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. Moreover, we introduce two additional supervision signals, a novel text-only blueprint schema and a 21-view image renderer that aids the agent's visual inspection, that better align the generation loop with how engineers iterate in practice. On S2O and Fusion360, 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. Together these signals move CAD programs toward artifacts that are not only visually plausible but also checked against physical and structural requirements.
Community
In this work, we introduce a more industry-native task formulation that requires a model to produce a fully assembled multi-part STEP file from a free-form engineering brief, which is then validated via finite element analysis (FEA).
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2605.17448 in a model README.md to link it from this page.
Cite arxiv.org/abs/2605.17448 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2605.17448 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.