Estimation and Correlation of Student Maturity with Social Attributes Using Large Language Models and Transformers

Abstract

Positive academic growth is foundational for a child's well-being and future success. While most children exhibit strong indicators of flourishing, early identification of potential challenges can significantly impact their development. This study utilizes AI-powered natural language processing techniques to assess a child's academic developmental progress based on classwork and conversational data. Analyzing homework samples from elementary and middle school students in the Capistrano Unified School District, we fine-tuned large language models to classify conversations by grade level. This classification identifies potential discrepancies between a child's developmental stage and conversational maturity. We found interesting correlations between social parameters like native languages, family income, race, number of siblings, and pets. Furthermore, we explored the potential of AI-driven interventions. Feedback from a system developed using Large Language Models helped students retain vocabulary and grammatical accuracy.