Conference proceeding
Sampling of Surfaces and Learning Functions in High Dimensions
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
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.
Details
- Title: Subtitle
- Sampling of Surfaces and Learning Functions in High Dimensions
- Creators
- Qing Zou - University of Iowa,Department of Mathematics,IA,USAMathews Jacob - University of Iowa,Department of Electrical and Computer Engineering,IA,USA
- Resource Type
- Conference proceeding
- Publication Details
- ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Vol.2020, pp.8354-8358
- DOI
- 10.1109/ICASSP40776.2020.9053876
- PMID
- 33603569
- NLM abbreviation
- Proc IEEE Int Conf Acoust Speech Signal Process
- ISSN
- 1520-6149
- eISSN
- 2379-190X
- Publisher
- IEEE
- Language
- English
- Date published
- 05/2020
- Academic Unit
- Roy J. Carver Department of Biomedical Engineering; Radiology; Electrical and Computer Engineering; Iowa Technology Institute; Iowa Neuroscience Institute; Radiation Oncology
- Record Identifier
- 9984070543902771
Metrics
15 Record Views