FairTrade: Achieving Pareto-Optimal Trade-Offs between Balanced Accuracy and Fairness in Federated Learning

Authors: Maryam Badar, Sandipan Sikdar, Wolfgang Nejdl, Marco Fisichella

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We provide empirical evidence of our framework’s efficacy through extensive experiments on five real-world datasets and comparisons with six baselines.
Researcher Affiliation Academia L3S Research Center, Leibniz University, Hannover, Germany
Pseudocode Yes Algorithm 1: Fair Trade server side algorithm
Open Source Code Yes For reproducibility, all resources associated with our research, including code and data, are publicly accessible at the provided repository link 1. 1https://github.com/badarm/Fair Trade
Open Datasets Yes We evaluate Fair Trade with five real-world datasets: (1) Bank (Bache and Lichman 2013), (2) Default (Bache and Lichman 2013), (3) Adult (Bache and Lichman 2013), (4) Law (Wightman 1998), and (5) KDD (Bache and Lichman 2013).
Dataset Splits No The paper describes how the dataset is distributed among clients (randomly or attribute-based) but does not provide specific training, validation, and test dataset split percentages or sample counts for model training.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts) used for running the experiments.
Software Dependencies No The paper mentions software components like BoTorch and Gaussian Process, but does not provide specific version numbers for any software dependencies.
Experiment Setup No The paper describes the model architecture and that some parameters (learning rate, regularization parameter) are optimized, but it does not provide specific hyperparameter values (e.g., initial learning rate, batch size, number of epochs, or specific values for 'no' and 'nc' from Algorithm 1) used for the reported experiments.