从USD场景到知识图谱：基于LLM的零样本本体接地
From USD Scenes to Knowledge Graphs: Zero-Shot Ontology Grounding with LLMs
AI总结：
研究利用大语言模型（LLM）零样本地将3D场景对象自动映射到本体类别，无需训练，在厨房场景中达到90-96%准确率，并揭示语义线索是关键。
摘要：
从3D仿真场景构建知识图谱对于机器人任务推理至关重要，但关键瓶颈——将场景对象接地到形式本体类别——仍然依赖于手工制作的字典，这些字典脆弱且无法跨资产泛化。我们研究大语言模型（LLM）是否能够自动化通用场景描述（USD）场景的接地步骤，作为一种零样本、无需训练的替代方案。在具有SOMA-HOME本体的厨房场景（125个对象）中，LLM在描述性名称下达到90-96%的精确匹配准确率，在缩写名称下达到49-89%，显著优于字典和嵌入基线。在完全不透明名称下，上下文增强提示可恢复高达48%的准确率。特征消融表明，LLM主要利用场景图中的语义线索（兄弟名称和父路径）；匿名化这些线索将准确率降至0-6%，而仅凭几何信息仅能达到4-17%。
英文摘要：
Constructing knowledge graphs from 3D simulation scenes is essential for robot task reasoning, but the key bottleneck, grounding scene objects to formal ontology classes, still relies on manually curated dictionaries that are brittle and do not generalize across assets. We investigate whether large language models (LLMs) can automate this grounding step for Universal Scene Description (USD) scenes as a zero-shot, training-free alternative. On a kitchen scene (125 objects) with SOMA-HOME Ontology, LLMs achieve 90-96% exact-match accuracy with descriptive names and 49-89% with abbreviated names, substantially outperforming dictionary and embedding baselines. Under fully opaque names, context-augmented prompting recovers up to 48%. Feature ablation reveals that LLMs primarily exploit semantic cues in the scene graph (sibling names and parent paths); anonymizing these cues reduces accuracy to 0-6%, while geometry alone yields only 4-17%.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
