College Graduation

Get In Touch

twitter twitter

Informatics West Room 233,
901 E. 10th Street
Bloomington, IN 47408

Current Research Projects

Epistatic SNP-SNP Interaction Using Information and Network Theory

Epistatic, or non-linear, interactions between SNPs are difficult to quantify. Systematically checking all combinations is both time-consuming and a computationally expensive problem. I will attempt to reproduce the results by Hu et al. in where they use information and network theory to investigate new associations betweens SNPs in glaucoma from the GLAUGEN dataset. I will then extend upon the methods to work with continuous phenotypes.

Quantifying Promoters using Nucleotide Diversity

Promoters are the start site of transcription. There are typically two shapes used to define promoters: broad and narrow. While the exact functional difference between the two are unknown, studies have shown the difference exists. Nucleotide Diveristy is a measure to what extent polymorphism exists in a popoulation. We use this measure to find patterns to differentiate promoter shapes.

Using Reinforcement Learning to Find Optimal Structured Interrupted Treatments in HIV

With such a high mutation rate, a constant antiviral treatment of HIV is not necessarily the best cure. A new method, Structured Treatment Interruptions, involves cycling treatments across different time intervals. Finding an optimal policy and time interval can be thought of as a reinforcement learning problem with the patient status as the state and the treatment as actions. Using Fitted Q Iteration, I plan on extending the work of Ernst et al. to find an optimal treatment and an optimal number of days to cycle the treatment.

Channel Decoding In Networks

Sending messages over a channel can lead to errors. Because of this, there are many ways to encode messages so that they can be decoded by the receiver with minimal error. Edge data from networks can also be sent as a message. We look to explore methods of efficiently passing edge data with minimal data loss.


HRL Laboratories

I interned at HRL Laboratories from May to August of 2016 in Malibu, California. I worked on building a new kind of a agent based model in order to mimic the posting behavior of protesters during times of social unrest. We validated our model using data about the Ferguson Protests of 2014 from a complete Tumblr dataset that included all blog posts and reblogs in that time period. During my time at HRL, I learned how to use Hadoop, how to work with high performance computers, and how to work in a company setting. This project was done under the guidance of Tsai-Ching Lu, Jiejun Xiu, and Aruna Jammalamadaka.

Past Research Projects

  • Easily Accessible Internet Security - HATS
  • Compartmental ODE Model of Cholera
  • Robustness of Zipf's Law in an evolving language
  • Opinion Volatility and the Stability of Democratic Communication Networks