FedCampus is a project that aims to implement a privacy-preserving data platform for the smart campus. This platform involves edge devices (e.g., smartphones, smartwatches, and IoTs), wireless networks (e.g., WIFI/4G/5G), and will provide insights into applications, such as human behavior, autonomous vehicle transportation, and smart healthcare. This project will benefit further research fields such as privacy-preserving machine learning, federated learning (FL), and privacy-preserving data analytics, federated analytics (FA).
The traditional computing model transfers data from local devices to a central server for processing. However, with the proliferation of edge devices such as smartphones and IoT devices, edge computing and distributed data management are becoming increasingly important. Sending a large amount of data from the source to the centralized data center is expensive. Using the edge computing model can reduce the amount of data sent from the site to the data center because end users only need to send critical data. Edge computing allows devices to store and process most of the data locally. Additionally, edge devices can encrypt local critical data before transferring private information to the central server for further processing.
This project aims to design a system for collecting and analyzing data from edge devices, including smartphones, smartwatches, and IoT devices. This system aims to provide valuable insights into the research areas of federated analytics (FA) and federated learning (FL). [Brief introduction of the system design, the benefits for research]
Some potential applications of this system include:
Federated analytics: By collecting and analyzing data from multiple devices, this system can support the research of federated analytics, which aims to gain insights from data that is distributed across different locations and devices. For example, in the healthcare industry, this system could be used to analyze patient data from multiple devices in a privacy-preserving manner. [DP algorithm example] Federated learning (FL): FL is a promising solution to solve user privacy problems regarding data sharing. It involves training machine learning models on data distributed across different devices and training statistical models over remote devices while keeping data localized. [Sleep efficiency example]