Edge Data

Edge data processing focuses on performing data analysis and machine learning closer to the data source, improving latency and bandwidth efficiency. Current research emphasizes optimizing model training and inference on resource-constrained edge devices, employing techniques like federated learning, sparse training, and personalized models (e.g., Bayesian Ridge Regression) to address data heterogeneity and scarcity. This approach is crucial for applications requiring real-time responses and enhanced privacy, impacting fields like healthcare (e.g., virus severity prediction) and vehicular networks, while also raising important security concerns addressed through techniques like data quarantine.