English

Learning SMaLL Predictors

Machine Learning 2018-03-08 v1

Abstract

We present a new machine learning technique for training small resource-constrained predictors. Our algorithm, the Sparse Multiprototype Linear Learner (SMaLL), is inspired by the classic machine learning problem of learning kk-DNF Boolean formulae. We present a formal derivation of our algorithm and demonstrate the benefits of our approach with a detailed empirical study.

Keywords

Cite

@article{arxiv.1803.02388,
  title  = {Learning SMaLL Predictors},
  author = {Vikas K. Garg and Ofer Dekel and Lin Xiao},
  journal= {arXiv preprint arXiv:1803.02388},
  year   = {2018}
}
R2 v1 2026-06-23T00:44:23.656Z