Research Charts Path for More Intelligent Wireless Communications

Cong Shen’s arrival at UVA’s School of Engineering coincides with the advent of 5G communication. “5G has created new challenges for communication theorists and practitioners to rethink system designs,” Shen said. The blending of machine learning and artificial intelligence with communication network theory—Shen’s specialty—can inform this thinking.

As an assistant professor in the Charles L. Brown Department of Electrical and Computer Engineering, Shen hopes to innovate deep learning theory and develop new tools to clear a path toward more intelligent wireless communications.

Today’s deep learning is very powerful, but it relies on large number of end users transmitting their private data that a service provider needs to train deep learning models. This reliance on data presents two problems. First, it is not privacy-friendly. Second, centralized processing leads to longer delays and requires large, centralized computing resources. Shen’s research shows that by adaptively partitioning deep neural networks into many pieces and allowing these partitioned networks to be trained at the end user’s device, both limitations can be overcome. Careful design of the partition strategy, to make the optimal tradeoff between performance and privacy for battery-powered mobile devices, is a focal point for Shen’s research at UVA.

In many ways, Shen helped usher in this new era while working for Qualcomm Research, SpiderCloud Wireless, Silvus Technologies and in various full time and consulting roles. Additionally, he was a professor in the School of Information Science and Technology at the University of Science and Technology of China amidst the boom in China’s communications industry. He credits his experience in Qualcomm’s research and development division for shifting his gaze toward real-world impacts and societal benefit. UVA Engineering empowers Shen to pursue his passion, to generate new knowledge and technologies that will drive modern life.