Dima Fayyad
The need to attain faster data rates and ultra-low latency are the driving forces behind emerging wireless standards and the promise of 5G communication. Many incredible, bleeding-edge applications, such as community/shared virtual reality experiences and self-driving cars, will crucially rely on the ubiquitous availability and real-time reconfigurability of high-speed wireless links.
The wireless channel and surrounding environment change rapidly, in the order of milliseconds. The current state of the art in wireless protocol design, involving multiple trials and sub-optimal selections with little room for self-optimization, is not geared to support the scale, heterogeneity, and real-time adaptation needed for these future applications.
Machine learning algorithms able to operate over unprocessed wireless signals from a multitude of transmission protocols have the potential of meeting both the latency requirements associated with these applications, as well as the adaptability necessary in the presence of an ever-changing, volatile wireless environment.
The Institute of Experiential AI has a strong presence and commitment to research over the applications of AI to wireless communications and IoT. In close collaboration with the Institute for the Wireless Internet of Things, EAI faculty participate in projects studying the application of deep learning algorithms to the wireless spectrum; in applications ranging from fingerprinting, spectrum allocation, and 5G networks. EAI faculty are developing real-time, robust adaptive learning algorithms to channel conditions while also meeting efficiency constraints imposed by wireless hardware limitations.