Acceleration for Compressed Gradient Descent in Distributed and Federated Optimization

Authors: Zhize Li, Dmitry Kovalev, Xun Qian, Peter Richtarik

ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we conduct several experiments on real-world datasets which corroborate our theoretical results and confirm the practical superiority of our accelerated methods.
Researcher Affiliation Academia 1King Abdullah University of Science and Technology, Thuwal, Kingdom of Saudi Arabia.
Pseudocode Yes Algorithm 1 Accelerated CGD (ACGD) and Algorithm 2 Accelerated DIANA (ADIANA)
Open Source Code No The paper does not provide an explicit statement or link to open-source code for the described methodology. No mention of code availability in supplementary materials or a repository link was found.
Open Datasets Yes In our experiments we use four standard datasets, namely, a5a, mushrooms, a9a and w6a from the LIBSVM library. Some of the experiments are provided in the appendix.
Dataset Splits No The paper mentions using 'standard datasets' from the LIBSVM library but does not explicitly provide specific train/validation/test dataset splits (e.g., percentages, sample counts, or references to predefined splits within the paper).
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU/GPU models, memory specifications, or cloud instance types) used for running its experiments. It only mentions 'The default number of nodes/machines is 20'.
Software Dependencies No The paper does not provide specific software dependencies (e.g., library names with version numbers) needed to replicate the experiment.
Experiment Setup Yes In our experiments, we use the theoretical stepsize and parameters for all the three algorithms: vanilla distributed compressed gradient descent (DCGD), DIANA (Mishchenko et al., 2019), and our ADIANA (Algorithm 2). The default number of nodes/machines is 20 and the regularization parameter λ = 10−3.