Prof. Soheil Mohajer has been awarded the CAREER award by the National Science Foundation’s Faculty Early Career Development (CAREER) Program. This is one of the most prestigious awards instituted by the NSF to recognize and support faculty early in their careers, who show the potential to “serve as academic role models in research and education and to lead advances in the mission of their department or organization.” The minimum CAREER award size is typically $500,000 (for the Directorate for Engineering) and is disbursed over a 5-year period.
The rising demand for broadband data has driven an increase in the volume of network traffic. And despite improvements in wireless communication technology, data rates continue to lag behind the exponential growth in demand. In such a scenario, opportunistic transmission strategies based on network characteristics, demand profile, and content type can help meet expectations.
Broadband video, a major contributor to this traffic explosion, is typified by repeat requests by multiple users, and having highly variant temporal traffic. These properties open up an opportunity: cache the data at local storage units closer to users during off-peak hours of the network, and thereby reduce network traffic at peak hours. Recent developments in coded caching offer a promising solution for high data rates.
Soheil’s CAREER award winning project addresses the practical challenges of caching in real world communication networks, and pursues a theoretical foundation for adopting caching as a universal resource in data delivery networks. Some of these challenges include substantial variations in caching gains in networks with real characteristics such as time-varying channels, asynchronous and delay sensitive requests, and absence of a central coordinator.
Soheil’s project promises coding techniques to overcome these barriers, and improve system performance. A rate-distortion theoretical framework will be developed to characterize the fundamental trade-off between cache size, delivery rate, and reconstruction quality, alongside efficient coding schemes to achieve this tradeoff. The project will also study the interaction between caching gains and spatial diversity, and the successful completion of this study will lead to optimum resource (cache size, power, and rate) allocation as well as transmission scheduling in non-homogenous multi-input-multi-output (MIMO) networks.
A final outcome of the project includes development of software to simulate caching techniques for a range of networks and applications, supporting both research and education. This particular part of the project will support Prof. Mohajer in his efforts to integrate his research into the graduate and undergraduate curricula, and expand the research process and outcomes to include the local K-12 student community through outreach.
Soheil earned his doctoral degree in 2010 from École Polytechnique Fédérale de Lausanne (EPFL). He was a postdoctoral scholar at Princeton University from 2010 to 2011, and a postdoctoral fellow at University of California, Berkeley, from 2011 to 2013. He joined the University of Minnesota in 2014, as a faculty member with the Department of Electrical and Computer Engineering. His research interests include information theory, distributed storage systems, data delivery networks, and statistical machine learning. He has authored over 13 journal papers and 60 conference papers. Learn more about his research here.