Browsing by Author "Gowtham, L."

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    EdgeFedNet: Edge Server Based Communication and Computation Efficient Federated Learning
    (Springer, 2025) Gowtham, L.; Annappa, B.; Sachin, D.N.
    Federated learning (FL) is a new learning framework for training machine learning and deep learning models using data spread over several edge devices. Edge devices like mobile phones and IoT devices have constraints on computational power, resources, and connectivity for training the model. Also, many model parameters will be exchanged while training the model, leading to high communication costs in FL when bandwidth is limited. This paper presents EdgeFedNet a new form of training the model in FL. The proposed method reduces the model parameters by pruning the model and restricts the communication between clients and the cloud server by implementing edge servers. An edge server near a set of clients forms a cluster and coordinates the FL training. The aggregated model updates from all the edge servers are sent to the cloud server, restricting the frequent communication between the clients and the cloud server. The experimental results exhibit a remarkable reduction in the number model parameters (up to 54%) and effectively address the communication overhead by reducing communication rounds by 59% compared to the baseline approach FedAvg. These enhancements are achieved without sacrificing accuracy, presenting promising implications for more efficient model parameter pruning and communication strategies. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.
  • No Thumbnail Available
    Item
    FedPruNet: Federated Learning Using Pruning Neural Network
    (Institute of Electrical and Electronics Engineers Inc., 2022) Gowtham, L.; Annappa, A.; Sachin, D.N.
    Federated Learning (FL) is a distributed form of training the machine learning and deep learning models on the data spread over heterogeneous edge devices. The global model at the server learns by aggregating local models sent by the edge devices, maintaining data privacy, and lowering communication costs by just communicating model updates. The edge devices on which the model gets trained usually will have limitations towards power resource, storage, computations to train the model. This paper address the computation overhead issue on the edge devices by presenting a new method named FedPruNet, which trains the model in edge devices using the neural network model pruning method. The proposed method successfully reduced the computation overhead on edge devices by pruning the model. Experimental results show that for the fixed number of communication rounds, the model parameters are pruned up to 41.35% and 65% on MNIST and CIFAR-10 datasets, respectively, without compromising the accuracy compared to training FL edge devices without pruning. © 2022 IEEE.