My research lies at the intersection of Bayesian inference, signal processing, and parallel computing. For my dissertation, I have developed and rigorously analyzed several new algorithms for performing Bayesian model comparison on parallel computer architectures. These algorithms have applications in acoustics and audio signal processing, image processing, and other signal processing domains. These techniques were developed with the goal of accelerating data analysis in these domains.
While at the University of Mississippi, most recently I worked on a small business technology transfer (STTR) grant with a defense contractor. The purpose of the grant was to develop machine learning methods for performing radar target identification for specialized targets. For this project, I developed theoretical foundations, algorithms, and software for implementing Bayesian model comparison in a machine learning classification framework for performing this target identification. For this grant I also prepared reports and made presentations to our sponsor. Previously, I have also worked on grants for an intermodal transportation network design project and with the Broadband Wireless Access and Applications Center (BWAC).
I have also worked on several projects related to Bayesian inference relating to acoustics applications, such as loudspeaker modeling and design and acoustic room mode analysis.
Currently, I am teaching myself more about artificial neural networks (ANN) and deep learning. I am particularly interested in the analysis of time series (such as audio signals) using these techniques. I am also interested in exploring the intersection of Bayesian inference with ANNs. The model comparison techniques developed in my dissertation could have an application in improving Bayesian-regularized artificial neural networks (BRANNs).