DistributedDistributionallyRobustOptimizationwith Non-ConvexObjectives

Centralized machine learning requires gathering the data to a particular server to train models which incurs high communication overhead [46] and suffersprivacyrisks[43]. Asaremedy,distributedmachine learning methods havebeenproposed. Considering a distributed system composed ofN workers (devices), we denote the dataset of these workers as{D1,,DN}.