发表机构
Wadhwani School of Data Science and Artificial Intelligence, IIT Madras; Department of Mathematics, IIT Madras; Robert Bosch Center for Data Science and Artificial Intelligence, IIT Madras; Department of Data Science and Artificial Intelligence, IIT Madras(瓦德万纳数据科学与人工智能学院,印度理工学院马德拉斯分校; 数学系,印度理工学院马德拉斯分校; 罗伯特·博世数据科学与人工智能中心,印度理工学院马德拉斯分校; 数据科学与人工智能系,印度理工学院马德拉斯分校)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
提出LoCCA框架,通过节点扰动稳定性理论将切比雪夫节点映射到物理邻居,实现核矩阵分解的稳定高效逼近,在不规则几何上提供严格误差控制并加速压缩。
AI 中文摘要
我们提出了局部切比雪夫交叉逼近(LoCCA),这是一个数据驱动的框架,将光滑多项式插值与灵活的矩阵分解相结合。LoCCA动态地将理想网格节点映射到非结构化点集中最近的物理邻居。我们通过新的扰动理论支持该框架,证明将切比雪夫节点位移到物理数据点上仍能保持数学稳定性和精度,而不会产生爆炸性误差增长。在数值上,LoCCA即使在标准基于网格的方法失效的不规则几何上也能提供严格、可靠的误差控制,实现显著的加速和接近最优的矩阵压缩。
英文摘要
We propose Local Chebyshev Cross-Approximation (LoCCA), a data-driven framework that bridges smooth polynomial interpolation with flexible matrix factorizations. LoCCA dynamically maps ideal grid nodes to their nearest physical neighbors within unstructured point sets. We support this framework with a new perturbation theory, proving that displacing Chebyshev nodes onto physical data points retains mathematical stability and accuracy without explosive error growth. Numerically, LoCCA provides strict, reliable error control even on irregular geometries where standard grid-based methods fail, achieving substantial speedups and near-optimal matrix compression.
Comments26 pages, 12 figures, 5 tables