Federated learning is the new tide that is being associated with machine learning territory. It is an attempt to enable smart edge devices to confederate a mutual prediction model while the training data is residing at the respective edge device. This facilitates our data to be more secure, use less bandwidth, lower latency, and power consumption. We exercise the concept of federated learning in our neural network autoencoder model to detect the anomaly. Anomaly detection is finding the unusual pattern in a given data stream which may be a false or mal-entry in the pool of transactions. It helps us to prevent many online theft and scams which are detected using state-of-the-art machine learning and deep learning algorithms. All this has to be implemented in smart edge devices that have enough computing power to train the models provided to them.
Shubham Singh hasn't uploaded this paper.
Let Shubham know you want this paper to be uploaded.