LAG: Lazily Aggregated Gradient for Communication-Efficient Distributed Learning

Authors: Tianyi Chen, Georgios Giannakis, Tao Sun, Wotao Yin

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Numerical experiments on both synthetic and real data corroborate a significant communication reduction compared to alternatives.
Researcher Affiliation Academia University of Minnesota Twin Cities, Minneapolis, MN 55455, USA National University of Defense Technology, Changsha, Hunan 410073, China University of California Los Angeles, Los Angeles, CA 90095, USA
Pseudocode Yes Algorithm 1 LAG-WK, Algorithm 2 LAG-PS
Open Source Code No The paper does not provide an explicit statement or link to open-source code for the described methodology.
Open Datasets Yes The paper states: "Performance is also tested on the real datasets [2]: a) linear regression using Housing, Body fat, Abalone datasets; and, b) logistic regression using Ionosphere, Adult, Derm datasets". It also cites "M. Lichman, UCI machine learning repository, 2013. [Online]. Available: http://archive.ics.uci.edu/ml" as [36], which is a well-known public repository.
Dataset Splits No The paper mentions "Each dataset is evenly split into three workers" but does not provide specific training, validation, or test dataset split percentages, counts, or methodology.
Hardware Specification Yes All experiments were performed using MATLAB on an Intel CPU @ 3.4 GHz (32 GB RAM) desktop.
Software Dependencies No The paper states "All experiments were performed using MATLAB" but does not specify a version number for MATLAB or any other software dependencies with version numbers.
Experiment Setup Yes Stepsizes for LAG-WK, LAG-PS, and GD are chosen as α = 1/L; to optimize performance and guarantee stability, α = 1/(ML) is used in Cyc-IAG and Num-IAG. For LAG-WK, we choose ξd = ξ = 1/D with D = 10, and for LAG-PS, we choose more aggressive ξd = ξ = 10/D with D = 10. For logistic regression, the regularization parameter is set to λ = 10−3.