Lookahead Converges to Stationary Points of Smooth Non-convex Functions

Jianyu Wang, Vinayak Tantia, Nicolas Ballas, Michael RabbatICASSP 2020 · OpenReview

Abstract

The Lookahead optimizer [Zhang et al., 2019] was recently proposed and demonstrated to improve performance of stochastic first-order methods for training deep neural networks. Lookahead can be viewed as a two time-scale algorithm, where the fast dynamics (inner optimizer) determine a search direction and the slow dynamics (outer optimizer) perform updates by moving along this direction. We prove that, with appropriate choice of step-sizes, Lookahead converges to a stationary point of smooth non-convex functions. Although Lookahead is described and implemented as a serial algorithm, our analysis is based on viewing Lookahead as a multi-agent optimization method with two agents communicating periodically.
BibTeX
@inproceedings{wang2020lookahead,
  author = {Jianyu Wang and Vinayak Tantia and Nicolas Ballas and Michael Rabbat},
  title = {{Lookahead Converges to Stationary Points of Smooth Non-convex Functions}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {2020},
  url = {https://openreview.net/forum?id=FCGgjE_xUGp},
}

Loading related content...

Figure