Newest papers

VoxProg: 3D Spatial Reasoning through Voxel Programming

Zhichao Zhai, Menglei Zhao, Guikun Chen, Zihao Yan, Kanzhi Wu, Wenguan Wang, Jun Xiao · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.30127 · cited by 0 works (OpenAlex) · open-access PDF

Indonesian

Kemampuan model bahasa untuk menafsirkan tugas dan menghasilkan program menawarkan dasar yang menjanjikan untuk pertimbangan spasial 3D. Namun, mengikuti instruksi spasial dalam adegan 3D membutuhkan program untuk membangun referensi geometris yang bahasa meninggalkan implisit. Dalam pekerjaan ini, kami memperkenalkan VoxProg, yang merumuskan penempatan objek 3D melalui pemrograman voxel: sebuah model bahasa menerjemahkan interpretasi tugas ke dalam program yang dapat dieksekusi melalui pandangan voxel lokal dari permukaan adegan. Karena pandangan ini memungkinkan geometri permukaan indeks pada grid reguler dengan koordinat eksplisit dan hubungan lingkungan, sebuah program bukti dapat mengisolasi wilayah yang relevan dengan tugas dan batas-batas secara langsung dari adegan mesh, dan sebuah program pos downstream dapat mengukur geometri yang dihasilkan untuk menghitung target $SE($3) pose dan eksekusi umpansi

Machine translation (Meta M2M100); the English original follows.

Abstract

Language models' abilities to interpret tasks and generate programs offer a promising basis for 3D spatial reasoning. Yet following spatial instructions in 3D scenes requires programs to construct geometric references that language leaves implicit. In this work, we introduce VoxProg, which formulates 3D object placement through voxel programming: a language model translates task interpretations into executable programs over local voxel views of scene surfaces. Because these views index surface geometry on regular grids with explicit coordinates and neighborhood relations, an evidence program can isolate task-relevant regions and boundaries directly from scene meshes, and a downstream pose program can measure the resulting geometry to compute a target $SE(3)$ pose. Visual and numerical feedback from execution lets the model inspect intermediate selections and candidate placements to iteratively revise its code. To evaluate open-ended placement across functional use, geometric fit, and contextual arrangements, we introduce the 71-task VoxProg-Bench benchmark along with an evaluation suite pairing perceptual Intent Alignment and 3D Physical Validity. Across VoxProg-Bench and external benchmarks, VoxProg consistently outperforms relation-based, keypoint-based, and mesh-based baselines.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Voxel, Artificial intelligence, Suite, Executable, Task (project management)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange