K-means in space: a radiation sensitivity evaluation

Authors:
Kiri L. Wagstaff;Benjamin Bornstein
Affiliations:
California Institute of Technology, Pasadena, CA;California Institute of Technology, Pasadena, CA
Venue:
ICML '09 Proceedings of the 26th Annual International Conference on Machine Learning
Year:
2009

Citing 9
Cited 2

Cluster detection in background noise

Pattern Recognition
Accelerating exact k-means algorithms with geometric reasoning

KDD '99 Proceedings of the fifth ACM SIGKDD international conference on Knowledge discovery and data mining
Computing nearest neighbors for moving points and applications to clustering

Proceedings of the tenth annual ACM-SIAM symposium on Discrete algorithms
An empirical comparison of four initialization methods for the K-Means algorithm

Pattern Recognition Letters
An Efficient k-Means Clustering Algorithm: Analysis and Implementation

IEEE Transactions on Pattern Analysis and Machine Intelligence
Refining Initial Points for K-Means Clustering

ICML '98 Proceedings of the Fifteenth International Conference on Machine Learning
Valgrind: a framework for heavyweight dynamic binary instrumentation

Proceedings of the 2007 ACM SIGPLAN conference on Programming language design and implementation
Clustering data with measurement errors

Computational Statistics & Data Analysis
Tests and tolerances for high-performance software-implemehted fault detection

IEEE Transactions on Computers

Machine learning in space: extending our reach

Machine Learning
Machine learning for science and society

Machine Learning

Quantified Score

Hi-index	0.00

Visualization

Abstract

Spacecraft increasingly employ onboard data analysis to inform further data collection and prioritization decisions. However, many spacecraft operate in high-radiation environments in which the reliability of dataintensive computation is not known. This paper presents the first study of radiation sensitivity for k-means clustering. Our key findings are 1) k-means data structures differ in sensitivity, which is not determined solely by the amount of memory exposed; 2) no special radiation protection is needed below a data-set-dependent radiation threshold, enabling the use of faster, smaller, and cheaper onboard memory; and 3) subsampling improves radiation tolerance slightly, but the use of kd-trees unfortunately reduces tolerance. Our conclusions can help tailor k-means for use in future high-radiation environments.