à£à£Ö±²¥Ðã

Jun Huang, assistant professor in South à£à£Ö±²¥Ðã State University's Jerome J. Lohr College of Engineering, has received a grant from the National Science Foundation to alleviate network efficiency and data privacy concerns related to wireless communications networks. 

The United States — and the rest of the world — is becoming increasingly reliant on wireless networks like 5G systems for countless aspects of modern life. Wireless networks are used for both complex things, like mapping software in cars, and "simple" things, like video streaming.

Another way to think about wireless networks' importance is this: if they disappeared tomorrow, modern life would become drastically less connected and less convenient.

With , or fifth generation, seemingly everything can be, and is, connected. From cars to houses to coffee shops to entire cities — it’s a major improvement from the four previous generations of wireless networks.

But 5G networks are reliant on artificial intelligence — deep learning models, specifically — to create their web of individual networks on the same 5G network system. This concept is called 5G network slicing and requires each network owner to contribute data to feed the common dataset used to train the centralized deep-learning models needed for the networks to work effectively.

The connectivity of 5G makes it more vulnerable to cyberattacks, which raises significant concerns about network efficiency and data privacy. 

South à£à£Ö±²¥Ðã State University assistant professor Jun Huang is conducting research aimed at resolving these concerns in 5G, and eventually 6G, networks. Huang has been investigating an approach called "federated learning" that may provide new levels of security and will better protect private data.

"Federated learning is a way to train machine learning models using data stored on multiple devices, like phones or vehicles, without moving the data to a central place," Huang said. "Instead, only the learning progress (like model updates) is shared. Federated learning keeps the data private and secure."

, according to Huang, allows devices, like cars and drones, to work together to train models without sharing sensitive data.

"This approach fits real-world needs, such as privacy in autonomous driving or IoT (Internet of Things) networks, while helping to improve efficiency and collaboration across devices," Huang explained.

Through a , Huang will begin bridging the gap between theory and practice in relation to federated learning. Specifically, Huang's work will advance mobile terminal collaboration, key technology for next-generation networking systems that enables efficient device-to-device communication, collaborative computing and seamless space-air-ground integration.

"The goal of my research is to design methods that make federated learning work better in real-world systems, like (unmanned aerial vehicle) swarms, connected vehicles or smart cities," Huang said. "I aim to solve problems like reducing communication costs, handling different types of data, and ensuring the system runs smoothly."

The results of this work are expected to improve network performance and provide crucial foundational understandings of these concepts that will be key to future next-generation networks.

Huang first became interested in this work two years ago while at Baylor University. He was learning how to use artificial intelligence in decentralized systems, like networks of vehicles and mobile terminals, when he came across the concept of federated learning.

"I was excited by how federated learning can balance privacy and efficiency, which are key challenges in these areas," Huang said, "Over time, this became a focus of my work, especially for applications like autonomous driving and smart cities."