Automating Diploma Title Generation: Applying Machine Learning Techniques
Alba Merdani, Nelda Kote, Kleda Ternova, Enida Sheme · 2024
The generation of diploma thesis titles is a creative and often challenging task that requires significant effort and domain knowledge. This paper addresses the need for an automated solution to generate innovative and relevant diploma thesis titles using machine learning techniques. By leveraging advanced text generation models, we aim to assist students and researchers in brainstorming and developing new research topics efficiently. We utilize a bilingual dataset comprising diploma titles in both English and Albanian, collected over the past five years. This dataset allows us to explore the effectiveness of the models across different languages and identify language-specific challenges. Additionally, we developed a simple Flask application that provides an Application Programming Interface endpoint for generating new diploma titles based on user inputs, demonstrating the practical applicability of our approach. This paper contributes to the field of natural language processing by showcasing the potential of machine learning models in automating creative tasks. It also provides insights into their comparative performance.