Thesis Defense Announcement for Minh (Matthew) Hong To – 3/21/25 at 10 AM
March 3, 2025
Thesis Title: A Fine-Tuned BERT Model for Improved Querying of the Unmanned Aerial System Integration Safety and Security Technology Ontology
When: March 21st (10:00 AM)
Where: Simrall 432 or Online (Link)
Candidate: Minh (Matthew) Hong To
Degree: Master of Science in Electrical & Computer Engineering
Committee Members: Dr. Samee Khan, Dr. Asad Malik, Dr. Yu Luo
Abstract
The use of unmanned aerial vehicles (UAS) in all industries is steadily increasing every year. To govern the use of UAS, the Federal Aviation Administration (FAA) seeks to provide a foundation of rules and regulations for UAS operation in the National Airspace System (NAS). The UAS Integration Safety and Security Technology Ontology (ISSTO) was developed using the Web Ontology Language (OWL) in 2023. In 2024, a query application was developed to search ISSTO for information about the safety and security of UAS operations. While the application is functional, the search results can be further fine-tuned to match what the user is looking for. The thesis proposes a fine-tuned BERT model for the query application. The model is fine-tuned using the data in ISSTO. It gives the application a better understanding of the taxonomy of the ontology. The integration of machine learning into the query application allows the program to better understand what the user is asking for. The developed model improves the accuracy and efficiency of the query application to become a better tool for UAS operators.