Doctor of Philosophy (PhD)


Computer Science

Document Type



Data in many research fields continues to grow in both size and complexity. For instance, recent technological advances have caused an increased throughput in data in various biological-related endeavors, such as DNA sequencing, molecular simulations, and medical imaging. In addition, the variance in the types of data (textual, signal, image, etc.) adds an additional complexity in analyzing the data. As such, there is a need for uniquely developed applications that cater towards the type of data. Several considerations must be made when attempting to create a tool for a particular dataset. First, we must consider the type of algorithm required for analyzing the data. Next, since the size and complexity of the data imposes high computation and memory requirements, it is important to select a proper hardware environment on which to build the application. By carefully both developing the algorithm and selecting the hardware, we can provide an effective environment in which to analyze huge amounts of highly complex data in a large-scale manner. In this dissertation, I go into detail regarding my applications using big data and deep learning techniques to analyze complex and large data. I investigate how big data frameworks, such as Hadoop, can be applied to problems such as large-scale molecular dynamics simulations. Following this, many popular deep learning frameworks are evaluated and compared to find those that suit certain hardware setups and deep learning models. Then, we explore an application of deep learning to a biomedical problem, namely ADHD diagnosis from fMRI data. Lastly, I demonstrate a framework for real-time and fine-grained vehicle detection and classification. With each of these works in this dissertation, a unique large-scale analysis algorithm or deep learning model is implemented that caters towards the problem and leverages specialized computing resources.



Committee Chair

Park, Seung-Jong