Research

During my undergraduate experience, I had the opportunity to contribute to multiple projects. Now, my research focus is on bridging artifical intelligence and information security.

Binary Visualization for Detecting Benign & Malicious Files

Using a machine learning model to train a neural network to be able to distinguish between malicious & benign PDFs through a binary visualization representation.

With Python, I created a program to web scrape multiple websites and gather benign & malicious files. The program would then upload each website to Binvis.io which would then convert the file into a visual representation by binary class.

The program would then train a neural network using TensorFlow and classify images by either benign or malicious tags. Testing included inputing user’s files to run through program and test against the model to ensure user’s digital safety. Program released a 93% accuracy rate.

Waste Drone

Created a flying drone that uses object detection to analyze waste around sidewalks and roadways to minimize littering.

Coded a Tello Drone using Python and a machine learning model which implemented object-detection in order to process video footage and detect trash items and trash type.

The program stored a digital log for users to view photos of trash and time trash was detected. It utilizes YOLOv5 algorithm from Ultralytics and is trained with custom annotated data via Roboflow.io and Kaggle.com datasets. We used a Tello SDK to automate drone control.

Our main purpose was to create a prototype to be able to detect waste in order to further develop the detection to record car license plates whose passengers were seen littering.

Hopper’s Fables

Programmed child’s video game website which helps children develop mathematical, reading, and computational skills.

Expanding on video game website which researched the ability to strengthen math skills through computational thinking for students. The project was developed using JavaScript and allowed users to interact with website with the Blockly coding language. Targeted age group was 8-9.

Abstract was published and presented at the ACM MidSoutheast 2022.