ENHANCING PRIVACY AND ACCURACY IN FEDERATED LEARNING FOR REGRESSION WITH SERVER-SIDE FILTERING TO ADDRESS OUTLIERS

Abstract

In the contemporary landscape characterized by extensive data proliferation, the amalgamation of information derived from a multitude of devices is imperative for the  advanced machine learning models. Nevertheless, the centralization of such data engenders significant apprehensions regarding privacy, particularly when the data is fetched from a heterogeneous array of devices including mobile phones, cameras, sensors, computers, and workstations. Federated Learning proffers a solution to these privacy-related dilemmas by maintaining a decentralized architecture, thereby enabling local devices to preserve their data while concurrently exchanging model parameters. Despite its promise, Federated Learning encounters substantial obstacles concerning data quality, which may arise from inherent biases, the presence of outliers, and the utilization of compromised devices. To mitigate these challenges, we advocate for the implementation of a server-side filtering methodology within Federated Learning, specifically tailored for regression-related problems. Based on this architecture, local devices train the model on their own data sets and then send the learned parameters to a central server. The server is then tasked with the filtration of erroneous contributions, thereby enhancing the overall accuracy of the model. This methodology is substantiated through the application of the Mean Squared Error  metric, a widely recognized standard within regression analysis, thereby augmenting both the efficiency and dependability of the learning process while safeguarding user privacy an essential component of Federated Learning.