The National Science Foundation (NSF) and Intel are partnering to fund research for machine learning (ML) in wireless networks in hopes of accelerating a new wireless architecture that can keep up with modern demands.
“The wireless networks of the future need to support much higher requirements than what current wireless networks can deliver, and they also need to be secure and energy-efficient,” Margaret Martonosi, assistant director for Computer and Information Science and Engineering at NSF, said. “That is why NSF and Intel have contributed $9 million to advance research activities addressing some of the most challenging issues in the development of future wireless systems.”
The Machine Learning for Wireless Networking Systems (MLWiNS) program will award 10 to 15 grants between $300,000 and $1,500,000 on projects around ML for wireless, ML for spectrum management, and distributed ML over wireless edge networks. Through broad research on wireless-specific machine learning, NSF wrote in its announcement, the agency aims to enable new wireless architectures and systems for future applications.
“The fast time-varying statistics of wireless channels and the dynamics of diverse network traffic require ML approaches that can operate online and adapt system parameters in real time,” the program description explains. “The data sets available for adaptation are typically limited, thus warranting the combination of data-driven ML methods with classical model-based approaches for efficient parameter estimation.”