AWARE, Beyond Sentence Boundaries: A Contextual Transformer Framework for Identifying Cultural Capital in STEM Narratives

The newly developed framework AWARE is designed to improve the identification of cultural capital themes within STEM narratives by analyzing context beyond individual sentence boundaries. Traditional natural language processing (NLP) models often examine sentences in isolation, which limits their ability to detect cultural capital that is embedded subtly across student reflections rather than appearing as explicit keywords. AWARE addresses this limitation by employing a contextual transformer approach, enabling a more nuanced understanding of how cultural capital manifests in educational narratives. The framework aims to support equitable learning environments by better recognizing diverse cultural expressions in STEM education. This approach reflects a growing recognition that cultural capital in student narratives is complex and distributed, requiring advanced methods to capture its presence effectively. The development of AWARE aligns with ongoing efforts to enhance inclusivity and fairness in educational assessment through improved computational tools. Overall, AWARE represents a significant step forward in leveraging AI to understand and support cultural diversity in STEM learning contexts.