Scalable Non-linear Learning with Adaptive Polynomial Expansions

Part of Advances in Neural Information Processing Systems 27 (NIPS 2014)

Bibtex Metadata Paper Reviews Supplemental

Authors

Alekh Agarwal, Alina Beygelzimer, Daniel J. Hsu, John Langford, Matus J. Telgarsky

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.