Kelum Gajamannage, Ph.D.
I am an Assistant Professor (Tenure-Track) in the Department of Mathematics and Applied Mathematical Sciences at University of Rhode Island, Kingston, USA. My research interests encompass the invention of mathematical and statistical algorithms/methods for Machine Learning (ML), Computer Vision (CV), and Network Science. Especially, I am fascinated with the invention of algebraic and neural network dimensionality reduction algorithms for high-dimensional and big data. I enhance the mathematical foundation to facilitate novel CV discoveries in the domains of image/trajectory restoration, imputation, and inpainting. I recover partially observed/corrupted social, sensor, and computer networks with the aid of ML, and then analyzes networks’ characteristics. I explore the hidden low-dimensionality of the U.S. financial system using ML to forecast market crashes/recoveries. I am excited to get involved as an advisor for students by directing undergraduate and graduate research projects. He is an enthusiastic educator of Mathematics, Statistics, and Data Science related courses for undergraduates and graduates.
I was an Assistant Professor (Tenure-Track) of Mathematics from 2019 to 2023 in the Department of Mathematics & Statistics at Texas A&M University, Corpus Christi, USA. I was a Postdoctoral Scholar from 2016 to 2019 in the Department of Mathematical Sciences at Worcester Polytechnic Institute, USA. I graduated with a Ph.D. in Mathematics in 2016 from Clarkson University, USA. I was awarded an MS in Applied Statistics and a BS (with honor) in Mathematics by the University of Peradeniya, Sri Lanka.
My research interests encompass the invention of mathematical and statistical algorithms/methods for Machine Learning, Computer Vision, and Network Science. I am interested in developing novel Nonlinear Dimensionality Reduction (NDR) algorithms to assure geometry and smoothness of the underlying manifold of high-dimensional data. I use low-rank Matrix Completion (MC) techniques to recover corrupted distance matrices produced by partially observed data. I utilize MC techniques to capture and characterize topology and utilize NDR to learn temporal dimensionality of large market movements to detect perturbations in the U.S. financial system. I produce Artificial Neural Networks frameworks to reconstruct fragmented trajectories of collective motion and predict U.S. financial markets. I am also interested in formulating patch-based image denoising methods using eigenvectors of the geodesics’ Gramian matrix.
I have been teaching Mathematics, Statistics, and Data Science related classes since 2007. I have been teaching classes at Texas A&M University--Corpus Christi since Fall-2019. Before that, I had undergone diverse teaching assignments as follows: 3 years in Worcester Polytechnic Institute (WPI), USA; 3 years in Clarkson University, USA; 4 years in both Uva Wellassa University and University of Peradeniya, in Sri Lanka.
I utilize classroom technology such as PDF Annotator, Echo360, etc, in my teaching and enjoy teaching online classes apart from regular classes. I engaged with curriculum development and decision making as a Faculty Board Member and Course Committee Member at Uva Wellassa University, Sri Lanka.
