English

Improving Model Training by Periodic Sampling over Weight Distributions

Machine Learning 2020-03-23 v2 Computer Vision and Pattern Recognition Machine Learning

Abstract

In this paper, we explore techniques centered around periodic sampling of model weights that provide convergence improvements on gradient update methods (vanilla \acs{SGD}, Momentum, Adam) for a variety of vision problems (classification, detection, segmentation). Importantly, our algorithms provide better, faster and more robust convergence and training performance with only a slight increase in computation time. Our techniques are independent of the neural network model, gradient optimization methods or existing optimal training policies and converge in a less volatile fashion with performance improvements that are approximately monotonic. We conduct a variety of experiments to quantify these improvements and identify scenarios where these techniques could be more useful.

Keywords

Cite

@article{arxiv.1905.05774,
  title  = {Improving Model Training by Periodic Sampling over Weight Distributions},
  author = {Samarth Tripathi and Jiayi Liu and Unmesh Kurup and Mohak Shah and Sauptik Dhar},
  journal= {arXiv preprint arXiv:1905.05774},
  year   = {2020}
}
R2 v1 2026-06-23T09:06:30.661Z