reproducibilityindex.ai

Byzantine-Robust Federated Learning: Impact of Client Subsampling and Local Updates

Authors: Youssef Allouah, Sadegh Farhadkhani, Rachid Guerraoui, Nirupam Gupta, Rafael Pinot, Geovani Rizk, Sasha Voitovych

Reproducibility Variable	Result	LLM Response
Research Type	Experimental	We validate our theory by experiments on the FEMNIST and CIFAR-10 image classification tasks.
Researcher Affiliation	Academia	1EPFL, 2Sorbonne Universit e, LPSM 3University of Toronto.
Pseudocode	Yes	Algorithm 1 Fed Ro: Fed Avg with a robust aggregation rule A
Open Source Code	No	No explicit statement or link providing access to the authors' open-source code for their methodology was found.
Open Datasets	Yes	We use the FEMNIST dataset (Caldas et al., 2018) and CIFAR-10 dataset (Krizhevsky et al., 2009).
Dataset Splits	No	The paper mentions 'training error' and general FL concepts, but it does not specify explicit train/validation/test dataset splits (percentages, counts, or specific predefined splits) used for their experiments.
Hardware Specification	Yes	Machines used for all the experiments: 2 NVIDIA A10-24GB GPUs and 8 NVIDIA Titan X Maxwell 16GB GPUs.
Software Dependencies	No	The paper mentions the 'LEAF Library' and a shell script for data preprocessing but does not provide specific version numbers for any software dependencies.
Experiment Setup	Yes	We summarize the learning hyperparameters in Table 1. ... We list all the hyperparameters used for this experiment in Table 4. ... We list all the hyperparameters used for this experiment in Table 5.