期刊
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
卷 117, 期 33, 页码 19664-19669出版社
NATL ACAD SCIENCES
DOI: 10.1073/pnas.2001741117
关键词
stratification inference; singularities |; persistent cohomology
资金
- Engineering and Physical Sciences Research Council (EPSRC) [EP/R018472/1]
- EPSRC [EP/G037280/1, EP/K041096/1]
- Medical Research Council [EP/G037280/1]
- F. Hoffmann-La Roche AG
- Royal Society
- EPSRC [EP/R018472/1, EP/K041096/1] Funding Source: UKRI
The quest for low-dimensional models which approximate high -dimensional data is pervasive across the physical, natural, and social sciences. The dominant paradigm underlying most standard modeling techniques assumes that the data are concentrated near a single unknown manifold of relatively small intrinsic dimen-sion. Here, we present a systematic framework for detecting interfaces and related anomalies in data which may fail to sat-isfy the manifold hypothesis. By computing the local topology of small regions around each data point, we are able to par-tition a given dataset into disjoint classes, each of which can be individually approximated by a single manifold. Since these manifolds may have different intrinsic dimensions, local topol-ogy discovers singular regions in data even when none of the points have been sampled precisely from the singularities. We showcase this method by identifying the intersection of two sur-faces in the 24-dimensional space of cyclo-octane conformations and by locating all of the self-intersections of a Henneberg min-imal surface immersed in 3-dimensional space. Due to the local nature of the topological computations, the algorithmic burden of performing such data stratification is readily distributable across several processors.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据