Hi There! My name is Joshua Michalenko and I am a recent graduate from New Mexico State University's Electrical Engineering program and a 2nd year graduate student to Rice University's electrical engineering program. I've been interested in science, engineering and technology my whole life that I hope you will get to see on my site. My research interests are within deep learning, image and video processing, and Natural Language Processing, and probabilistic graphical models.
I'm a PhD student advised by Rich Baraniuk in the DSP group at Rice University, and I'm currently working on probabilistic frameworks for inferring misconceptions in textual responses to questions. Our group has two main focuses, developing new generative models that can explain the recent successes in deep learning and providing intelligent education resources to students in collaboration with the Rice founded startup OpenStax. My research lies in the middle of these areas and is focused on using Long-Short Term Memory (LSTM) Recurrent Neural Networks (RNN) in combination with probabilistic graphical models applied to question answering in natural language processing to infer if answers are correct, contain misconceptions, and even to track the ability of the respondent. I have previously worked on applying LSTMs to model mathematics derivations and student solutions to derivations as well.