通过合成数据挖掘3D模型的属性子空间，实现更高效精准的微调。
Mining Attribute Subspaces for Efficient Fine-tuning of 3D Foundation Models
- 构建可控变化的合成数据集，分别训练对应LoRA子空间。
- 发现不同属性子空间近似正交，可组合成压缩后的高效子空间。
- 仅用合成数据训练的子空间能泛化到真实数据，适合资源受限场景。
随着3D基础模型的兴起，针对下游任务的微调日益受到关注，其中LoRA是主流方法。由于3D数据在纹理、几何、相机运动和光照方面存在显著差异，提出三个基本问题：1）是否存在与每类变化相关的LoRA子空间？2）这些子空间是否解耦（即正交）？3）如何有效计算它们？本文给出解答：提出一种稳健方法，生成具有控制性变化的合成数据集，在每个数据集上微调一个LoRA适配器，并提取与各类变化对应的LoRA子空间。结果表明，这些子空间近似解耦。将它们整合后得到一个压缩的LoRA子空间，可实现更高效的微调并提升下游任务预测精度。特别地，该压缩子空间虽完全由合成数据训练得出，仍能泛化至真实数据集。消融实验验证了方法设计的有效性。
原文摘要 · Abstract (English)
With the emergence of 3D foundation models, there is growing interest in fine-tuning them for downstream tasks, where LoRA is the dominant fine-tuning paradigm. As 3D datasets exhibit distinct variations in texture, geometry, camera motion, and lighting, there are interesting fundamental questions: 1) Are there LoRA subspaces associated with each type of variation? 2) Are these subspaces disentangled (i.e., orthogonal to each other)? 3) How do we compute them effectively? This paper provides answers to all these questions. We introduce a robust approach that generates synthetic datasets with controlled variations, fine-tunes a LoRA adapter on each dataset, and extracts a LoRA sub-space associated with each type of variation. We show that these subspaces are approximately disentangled. Integrating them leads to a reduced LoRA subspace that enables efficient LoRA fine-tuning with improved prediction accuracy for downstream tasks. In particular, we show that such a reduced LoRA subspace, despite being derived entirely from synthetic data, generalizes to real datasets. An ablation study validates the effectiveness of the choices in our approach.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。