About Me
My name is Thomas Mercer, but I go by Ty. You can reach me at tmercer2022@my.fit.edu. I am an undergraduate at Florida Institute of Technology getting a B.S. in Computer Science. My degree is part of the Computer Engineering and Sciences department. As of Spring 2023, I am sophomore. I am set to graduate in either the Spring or Fall of 2025, so I am either still a sophomore or a junior starting in Fall 2023. I am aware that grade level is not typically dependent on start date, not end date, but I am a transfer student.
About My Mentor
Dr. Nasheen Nur is an Assistant Professor in the Department of Computer and Engineering Sciences at the Florida Institute of Technology (FIT). She completed her doctoral study at the University of North Carolina at Charlotte in 2021. Her research focuses on Natural Language Processing, Explainable Artificial Intelligence (XAI), AI in Education, and Accessible Computing. In her dissertation, she designed a software tool prototype for analyzing students’ behavior using temporal analytics and explainable artificial intelligence. Student advisers and data scientists collaborate iteratively to develop knowledge about student patterns of success and failure using this interactive tool. She specializes in concept-oriented visualization, machine learning models, and user interface techniques for explainable artificial intelligence (XAI). She was a Data Science Intern at the Pacific Northwest National Laboratory in 2018. She worked as a .NET front end developer at IQVIA Bangladesh after her undergraduate degree in Computer Science and Engineering from the Bangladesh University of Engineering & Technology.
About My Project
I am contributing to three separate projects as part of my DREU Project.
Project 1 — Machine Learning to Determine if Someone is Suicidal, to be Used in Suicide Prevention
The first and highest priority project involves training and evaluating a neural network designed to determine if a social media user did or will commit suicide based on their posts. For this desired outcome, the training data, where instances correspond to a social media post, must hold as attributes a boolean if the user did commit suicide, an anonymized user identifier. and the text of posts. The model generates a suicidality boolean that is compared against the boolean for if the person committed suicide.
People working on the project before me (and are continuing to work on it) have developed a model and are working to understand the rules used by the model to determine if a user is suicidal. One person is working on feeding both the posts from a user and the suicidality boolean from our model into a text generation model. From there, the text generation model should be able to provide reasoning behind the suicidality boolean. Another person is working on using layer integrated gradients to determine which words are considered most relevant by the model when determining the suicidality boolean.
The two main researchers described in the paragraph before this will be moving away from the project midway through the summer. In response, these researchers are planning to transfer knowledge over to us to continue research. This transition will occur the 6th week of my DREU program. As of now I am doing training with Python machine learning packages that will be necessary regardless of how I contribute. I am also contributing to the related work so that we can submit our research to publications and conferences.
Project 2 — Research to Assist People with ADHD and Their Managers in The Workplace
The second project is the creation of workshops to teach people with ADHD and managers of people with ADHD important skills to work with or accommodate ADHD. The workshops for ADHDers [people with ADHD] can teach ADHDers productivity and mindfulness skills, while the workshops for managers of ADHDers can teach managers how to interact with ADHDers healthily and understand ADHDers’ strengths and weaknesses, along with other topics for both workshops. Completing this project involves gathering research about accommodating and working with ADHD, collaborating with experts and people with ADHD, finding and preparing people to present at the workshops, and creating supplemental documents for the workshops.
This project is unlikely to be completed in full during the DREU Program and will be a continuing collaboration with Dr. Nur into further semesters.
Project 3 — Assisting With The Development of a Math Education App
Victor Tumbiolo is another student in the DREU Program with the same advisor as me, Dr. Nasheen Nur. As part of his DREU Program, he is working on an educational math app. I may contribute to this app and will blog about it if and when I do. He will have more complete information on the development of the app on his DREU Program website.