HG-CAD:面向计算机辅助设计的层级图学习用于材料预测与推荐
TLDR
HG-CAD通过层级图学习预测和推荐CAD装配体材料,在Fusion 360数据集上超越基线。
评分理由
The paper introduces a novel hierarchical graph representation for joint learning of body geometry and assembly topology, achieving strong results on material prediction. However, its scope is limited to material recommendation rather than broader CAD generation or reconstruction tasks, and evaluation is confined to a single dataset.
Read-first 评分解释
综合优先阅读分 27.9,由主题、引用、图谱、方法、可复现性和近期性等信号加权得到。 原始总分保留为 21。
研究版图角色
候选论文
排序敏感性
稳定性:volatile;排名波动范围:61。