Scalable Non-Linear Learning with Adaptive Polynomial Expansions

Abstract

Can we effectively learn a nonlinear representation in time comparable to linear learning? We describe a new algorithm that explicitly and adaptively expands higher-order interaction features over base linear representations. The algorithm is designed for extreme computational efficiency, and an extensive experimental study shows that its computation/prediction tradeoff ability compares very favorably against strong baselines.

Cite

Text

Agarwal et al. "Scalable Non-Linear Learning with Adaptive Polynomial Expansions." Neural Information Processing Systems, 2014.

Markdown

[Agarwal et al. "Scalable Non-Linear Learning with Adaptive Polynomial Expansions." Neural Information Processing Systems, 2014.](https://mlanthology.org/neurips/2014/agarwal2014neurips-scalable/)

BibTeX

@inproceedings{agarwal2014neurips-scalable,
  title     = {{Scalable Non-Linear Learning with Adaptive Polynomial Expansions}},
  author    = {Agarwal, Alekh and Beygelzimer, Alina and Hsu, Daniel J. and Langford, John and Telgarsky, Matus J},
  booktitle = {Neural Information Processing Systems},
  year      = {2014},
  pages     = {2051-2059},
  url       = {https://mlanthology.org/neurips/2014/agarwal2014neurips-scalable/}
}