Reproduce “FedMaJ: Ordinal Meta-Aggregation for Federated Learning Using Majority-Judgment Voting” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Driss Choukri, Essaid Sabir, Halima Elbiaze, Abdelkrim Haqiq
FedMaJ selects among fixed aggregation rules using client grades, choosing the rule with the highest median at voting rounds. In a synthetic OTA-FL proxy experiment, it achieved proxy accuracy of 0.9095 ± 0.0259 and Jain influence fairness of 0.9535 ± 0.0106, versus FedAvg’s 0.9294 ± 0.0257 and 0.8871 ± 0.0158. Under tested extreme evaluations, majority judgment changed rules less often than average-score selection.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.