Federated Distributionally Robust Optimization for Phase Configuration\n of RISs

In this article, we study the problem of robust reconfigurable intelligent\nsurface (RIS)-aided downlink communication over heterogeneous RIS types in the\nsupervised learning setting. By modeling downlink communication over\nheterogeneous RIS designs as different workers that learn how to optimize phase\nconfigurations in a distributed manner, we solve this distributed learning\nproblem using a distributionally robust formulation in a\ncommunication-efficient manner, while establishing its rate of convergence. By\ndoing so, we ensure that the global model performance of the worst-case worker\nis close to the performance of other workers. Simulation results show that our\nproposed algorithm requires fewer communication rounds (about 50% lesser) to\nachieve the same worst-case distribution test accuracy compared to competitive\nbaselines.\n