Decentralized gradient methods: does topology matter?

Abstract : Consensus-based distributed optimization methods have recently been advocated as alternatives to parameter server and ring all-reduce paradigms for large scale training of machine learning models. In this case, each worker maintains a local estimate of the optimal parameter vector and iteratively updates it by averaging the estimates obtained from its neighbors, and applying a correction on the basis of its local dataset. While theoretical results suggest that worker communication topology should have strong impact on the number of epochs needed to converge, previous experiments have shown the opposite conclusion. This paper sheds lights on this apparent contradiction and show how sparse topologies can lead to faster convergence even in the absence of communication delays.
Complete list of metadatas

https://hal.inria.fr/hal-02430485
Contributor : Giovanni Neglia <>
Submitted on : Tuesday, January 7, 2020 - 12:54:45 PM
Last modification on : Wednesday, January 22, 2020 - 4:43:17 PM

File

AISTATS2020.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02430485, version 1

Collections

Citation

Giovanni Neglia, Chuan Xu, Don Towsley, Gianmarco Calbi. Decentralized gradient methods: does topology matter?. AISTATS 2020 - 23rd International Conference on Artificial Intelligence and Statistics, Jun 2020, Palermo, Italy. ⟨hal-02430485⟩

Share

Metrics

Record views

56

Files downloads

349