EE Systems Seminar
Abstract: In the year 1918, Indian mathematician Srinivasa Ramanujan proposed a set of sequences called Ramanujan Sums as bases to expand arithmetic functions in number theory. Today, almost 100 years later, we will show that these sequences re-emerge as beautiful tools in a completely different context: For the extraction of periodic patterns in data. Combined with modern DSP techniques such as compressed sensing, Ramanujan Sums can be forged into powerful algorithms for numerous applications. This talk will demonstrate these using examples that include repeats in DNA and protein sequences.
From a theoretical perspective too, Ramanujan sums reveal a number of previously unknown but fundamental properties of period estimation, such as bounds on the absolute minimum number of samples required for unambiguous period estimation.
Bio: Srikanth V. Tenneti is currently a PhD student in the Digital Signal Processing Group of Prof. P. P. Vaidyanathan at Caltech. He received his B.Tech. degree in EE from the Indian Institute of Technology - Bombay, India, in 2012, and his M.S. degree in EE from Caltech in 2014. His current research is focused on developing a new framework for periodicity analysis using a combination of ideas from classical number theory (Ramanujan Sums), combinatorics, compressive sensing, and filter design. His work based on Ramanujan sums received a student paper award at the Asilomar Conference on Signals, Systems and Computers in 2015.
Host: Victoria Kostina