Lookahead Optimizer: k steps forward, 1 step back

Michael Ruogu Zhang,James Lucas,Geoffrey E. Hinton,Jimmy Ba

Published 2019 in Neural Information Processing Systems

ABSTRACT

The vast majority of successful deep neural networks are trained using variants of stochastic gradient descent (SGD) algorithms. Recent attempts to improve SGD can be broadly categorized into two approaches: (1) adaptive learning rate schemes, such as AdaGrad and Adam, and (2) accelerated schemes, such as heavy-ball and Nesterov momentum. In this paper, we propose a new optimization algorithm, Lookahead, that is orthogonal to these previous approaches and iteratively updates two sets of weights. Intuitively, the algorithm chooses a search direction by looking ahead at the sequence of fast weights generated by another optimizer. We show that Lookahead improves the learning stability and lowers the variance of its inner optimizer with negligible computation and memory cost. We empirically demonstrate Lookahead can significantly improve the performance of SGD and Adam, even with their default hyperparameter settings on ImageNet, CIFAR-10/100, neural machine translation, and Penn Treebank.

PUBLICATION RECORD

  • Publication year

    2019

  • Venue

    Neural Information Processing Systems

  • Publication date

    2019-07-19

  • Fields of study

    Mathematics, Computer Science

  • Identifiers
  • External record

    Open on Semantic Scholar

  • Source metadata

    Semantic Scholar

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-51 of 51 references · Page 1 of 1

CITED BY

Showing 1-100 of 830 citing papers · Page 1 of 9