Research map / Point clouds / Shape and surface completion
ReACT: Reward-informed Autoregressive Decision CAD Transformer
Yijie Ding, Yang Liu, Haobo Jiang, Jianmin Zheng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add return-to-go reward conditioning (ablation) | ReACT · DeepCAD (ablation) | mean CD (x10^3): 9.967 -> 3.929 | Improved |
| add global alignment reward Ralign (ablation) | ReACT · DeepCAD (ablation) | mean CD (x10^3): 8.123 -> 3.929 | Improved |
| add LBP coverage reward Rlbp (ablation) | ReACT · DeepCAD (ablation) | sketch F1: 67.49 -> 75.76 | Improved |
| add time-efficiency reward Rtime (ablation) | ReACT · DeepCAD (ablation) | invalid rate: 4.14 -> 2.68 | Improved |
| add scaffold-aware encoding (ablation) | ReACT · DeepCAD (ablation) | mean CD (x10^3): 6.110 -> 3.929 | Improved |
| use ReACT instead of Text2CAD* imitation baseline | ReACT · DeepCAD | mean CD (x10^3): 9.07 -> 3.92 | Improved |
| evaluate cross-dataset on Fusion360 | ReACT · Fusion360 (cross-dataset) | median CD (x10^3): 1.79 -> 0.66 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TransCAD — TransCAD: A Hierarchical Transformer for CAD Sequence Inference from Point Clouds (ECCV 2024) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- cadrille: Multi-modal CAD Reconstruction with Reinforcement Learning (ICLR 2026 · similarity 0.89)
- StepCAD: Mesh-to-CAD Code Generation via LLM Policy and Geometry-Guided Search (NeurIPS 2026 · similarity 0.88)
- CAD-Coder: Text-to-CAD Generation with Chain-of-Thought and Geometric Reward (NeurIPS 2025 · similarity 0.87)
- ReCAD: Reinforcement Learning Enhanced Parametric CAD Model Generation with Vision-Language Models (AAAI 2026 · similarity 0.87)
- CAD-SIGNet: CAD Language Inference from Point Clouds using Layer-wise Sketch Instance Guided Attention (CVPR 2024 · similarity 0.86)
- From Intent to Execution: Multimodal Chain-of-Thought Reinforcement Learning for Precise CAD Code Generation (AAAI 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
