First-Order Softmax Weighted Switching Gradient Method for Distributed Stochastic Minimax Optimization with Stochastic Constraints

First-Order Softmax Weighted Switching Gradient Method for Distributed Stochastic Minimax Optimization with Stochastic Constraints explores A federated learning algorithm that optimizes worst-case client performance with a novel softmax-weighted switching gradient method, offering a stable alternative to primal-dual approaches.. Commercial viability score: 5/10 in Federated Learning.