Learning a Set of Directions
Abstract
Assume our data consists of unit vectors (directions) and we are to find a small orthogonal set of the “the most important directions” summarizing the data. We develop online algorithms for this type of problem. The techniques used are similar to Principal Component Analysis which finds the most important small rank subspace of the data.The new problem is significantly more complex since the online algorithm maintains uncertainty over the most relevant subspace as well as directional information.
Cite
Text
Koolen et al. "Learning a Set of Directions." Annual Conference on Computational Learning Theory, 2013.Markdown
[Koolen et al. "Learning a Set of Directions." Annual Conference on Computational Learning Theory, 2013.](https://mlanthology.org/colt/2013/koolen2013colt-learning/)BibTeX
@inproceedings{koolen2013colt-learning,
title = {{Learning a Set of Directions}},
author = {Koolen, Wouter M. and Nie, Jiazhong and Warmuth, Manfred K.},
booktitle = {Annual Conference on Computational Learning Theory},
year = {2013},
pages = {851-866},
url = {https://mlanthology.org/colt/2013/koolen2013colt-learning/}
}