Logo image
Sampling of Surfaces and Learning Functions in High Dimensions
Conference proceeding

Sampling of Surfaces and Learning Functions in High Dimensions

Qing Zou and Mathews Jacob
ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Vol.2020, pp.8354-8358
05/2020
DOI: 10.1109/ICASSP40776.2020.9053876
PMID: 33603569
url
https://www.ncbi.nlm.nih.gov/pmc/articles/7885619View
Open Access

Abstract

The efficient representation of data in high-dimensional spaces is a key problem in several machine learning tasks. To capture the non-linear structure of the data, we model the data as points living on a smooth surface. We model the surface as the zero level-set of a bandlimited function. We show that this representation allows a non-linear lifting of the surface model, which will map the points to a low-dimensional subspace. This mapping between surfaces and the well-understood subspace model allows us to introduce novel algorithms (a) to recover the surface from few of its samples and (b) to learn a multidimensional bandlimited function from training data. The utility of these algorithms is introduced in practical applications including image denoising.
Signal processing algorithms Training data union of surfaces Bandwidth Data models learning Kernel Task analysis Surface treatment

Details

Metrics

Logo image