A variety of modern AI products essentially require raw user data for training diverse machine learning models. With the increasing concern on data privacy, federated learning, a decentralized learning framework, enables privacy-preserving training of models by iteratively aggregating model updates from participants, instead of aggregating raw data. Since all the participants, i.e., mobile devices, need to transfer their local model updates concurrently and iteratively over mobile edge networks, the network is easily overloaded, leading to a high risk of transmission failures. Although previous works on transmission protocols have already tried their best to avoid transmission collisions, the number of iterative concurrent transmissions should be fundamentally decreased. Inspired by the fact that raw data are often generated unevenly among devices, those devices with a small proportion of data could be properly excluded since they have little effect on the convergence of models. To further guarantee the accuracy of models, we propose to properly select a subset of devices as participants to ensure the given proportion of involved data. Correspondingly, we propose to minimize the risk against the transmission failures during model updates. Afterwards, we design a randomized algorithm (ranRFL) to choose suitable participants by using a series of delicately calculated probabilities, and prove that the result is concentrated on its optimum with high probability. Extensive simulations show that through delicate participant selection, ranRFL decreases the maximal error rate of model updates by up to 38.3% compared with the state-of-the-art schemas.INDEX TERMS Concurrent transmissions, federated learning, mobile edge networks.

In order to determine the optimal structural parameters of a plastic centrifugal pump in the framework of an orthogonal-experiment approach, a numerical study of the related flow field has been performed using CFX. The thickness S, outlet angle β2, inlet angle β1, wrap angle, and inlet diameter D1 of the splitter blades have been considered as the variable factors, using the shaft power and efficiency of the pump as evaluation indices. Through a parametric analysis, the relative importance of the influence of each structural parameter on each evaluation index has been obtained, leading to the following combinations: β1 19°, β2 35°, S 2 mm, wrap angle 154°, and D1 85 mm (corresponding to the maximum efficiency of 75.48%); β1 19°, β2 20°, S 6 mm, wrap angle 158°, and D1 81 mm (corresponding to the minimum shaft power of 75.48%). Moreover, the grey correlation method has been applied to re-optimize the shaft power and efficiency of the pump, leading to the following optimal combination: β1 19°, β2 15°, S 4 mm, D1 81 mm, and wrap angle 152°(corresponding to the maximum efficiency of 71.81% and minimum shaft power of 2.187 kW).

No abstract

No abstract

No abstract