Speedy Greedy Feature Selection: Better Redshift Estimation via Massive Parallelism

Overview
Related items

Abstract:

Nearest neighbor models are among the most basic tools in machine learning, and recent work has demonstrated their effectiveness in the field of astronomy. The performance of these models crucially depends on the underlying metric, and in particular on the selection of a meaningful subset of informative features. The feature selection is task-dependent and usually very time-consuming. In this work, we propose an efficient parallel implementation of incremental feature selection for nearest neighbor models utilizing nowadays graphics processing units. Our framework provides significant computational speed-ups over its sequential single-core competitor of up to two orders of magnitude. We demonstrate the applicability of the overall scheme on one of the most challenging tasks in astronomy: redshift estimation for distant galaxies.

SEEK ID: https://publications.h-its.org/publications/446

Research Groups: Astroinformatics

Publication type: InProceedings

Journal: European Symposium on Artificial Neural Networks, Computational Intelligence And Machine Learning

Book Title: 22st European Symposium on Artificial Neural Networks, Computational Intelligence And Machine Learning Bruges April 23-24-25, 2014

Publisher: ESANN

Citation: European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. Bruges (Belgium), 23-25 April 2014

Date Published: 17th Mar 2014

Registered Mode: manually

Authors: F. Gieseke, Kai L. Polsterer, Cosmin Eugen Oancea, Christian Igel

Help

Tree Split Graph