Efficient Online Bootstrapping for Large Scale Learning
Machine Learning
2013-12-19 v1
Abstract
Bootstrapping is a useful technique for estimating the uncertainty of a predictor, for example, confidence intervals for prediction. It is typically used on small to moderate sized datasets, due to its high computation cost. This work describes a highly scalable online bootstrapping strategy, implemented inside Vowpal Wabbit, that is several times faster than traditional strategies. Our experiments indicate that, in addition to providing a black box-like method for estimating uncertainty, our implementation of online bootstrapping may also help to train models with better prediction performance due to model averaging.
Keywords
Cite
@article{arxiv.1312.5021,
title = {Efficient Online Bootstrapping for Large Scale Learning},
author = {Zhen Qin and Vaclav Petricek and Nikos Karampatziakis and Lihong Li and John Langford},
journal= {arXiv preprint arXiv:1312.5021},
year = {2013}
}
Comments
5 pages, appeared at Big Learning Workshop at Neural Information Processing Systems 2013