Leveraging Transformer-based LLMs to Understand Barriers and Enablers for Women in Engineering Education

In recent years, transformer-based Language Models (TLMs), most notably known as Fine-tuned Generative Pre-trained Transformers (GPT), have emerged as potent research tools in the social sciences. This study explores the obstacles and facilitators that women in engineering education encounter using a Fine-tuned GPT. By means of a complete corpus analysis that includes academic literature, internet forums, and institutional records, the research endeavors to reveal subtle obstacles impeding women's involvement and accomplishment in engineering fields. The GPT is pre-trained on a large dataset of engineering education texts as part of the technique, and it is then particularly tuned to understand language related to gender. With this method, the model can identify underlying patterns that qualitative methods can miss and produce relevant responses.Important discoveries highlight structural problems that affect women's experiences in engineering education, including widespread gender stereotypes, implicit biases, and a lack of mentorship opportunities. The study also finds important facilitators that help women's academic and professional success in engineering areas, such as encouraging learning environments, positive role models, and inclusive legislation. GPT research provides insights that can be used by educational institutions and policymakers to develop focused actions that promote gender equity in engineering education. This study adds to the growing body of research on using natural language processing methods to address societal issues, especially those pertaining to gender diversity in STEM professions. The model's ability to handle bigger and more varied datasets will be improved in the future, along with the exploration of new intersectionality dimensions and the long-term effects of interventions meant to improve inclusion in engineering education.