ICML 2021spotlight41 citations

Fast margin maximization via dual acceleration

Ziwei Ji, Nathan Srebro, Matus Telgarsky

Abstract

We present and analyze a momentum-based gradient method for training linear classifiers with an exponentially-tailed loss (e.g., the exponential or logistic loss), which maximizes the classification margin on separable data at a rate of O(1/t^2). This contrasts with a rate of O(1/log(t)) for standard gradient descent, and O(1/t) for normalized gradient descent. The momentum-based method is derived via the convex dual of the maximum-margin problem, and specifically by applying Nesterov acceleration to this dual, which manages to result in a simple and intuitive method in the primal. This dual view can also be used to derive a stochastic variant, which performs adaptive non-uniform sampling via the dual variables.

BibTeX
@InProceedings{pmlr-v139-ji21a,
  title = 	 {Fast margin maximization via dual acceleration},
  author =       {Ji, Ziwei and Srebro, Nathan and Telgarsky, Matus},
  booktitle = 	 {Proceedings of the 38th International Conference on Machine Learning},
  pages = 	 {4860--4869},
  year = 	 {2021},
  editor = 	 {Meila, Marina and Zhang, Tong},
  volume = 	 {139},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {18--24 Jul},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v139/ji21a/ji21a.pdf},
  url = 	 {https://proceedings.mlr.press/v139/ji21a.html},
  abstract = 	 {We present and analyze a momentum-based gradient method for training linear classifiers with an exponentially-tailed loss (e.g., the exponential or logistic loss), which maximizes the classification margin on separable data at a rate of O(1/t^2). This contrasts with a rate of O(1/log(t)) for standard gradient descent, and O(1/t) for normalized gradient descent. The momentum-based method is derived via the convex dual of the maximum-margin problem, and specifically by applying Nesterov acceleration to this dual, which manages to result in a simple and intuitive method in the primal. This dual view can also be used to derive a stochastic variant, which performs adaptive non-uniform sampling via the dual variables.}
}
Fast margin maximization via dual acceleration · ICML 2021