Maze Assessment
Gainesville, FL | February 2024 - Present
Principal Investigator Department
Dr. Walter Leite Education
Research Focus
This project aims to develop an AI-based stealth assessment to measure oral reading fluency (ORF) in elementary school students without traditional testing methods. The goal is to create features from audio recordings of students reading to help personalize learning and improve educational technologies. It seeks to identify key features from students’ reading practice, such as reading accuracy, rate, pauses, and word difficulty, that can predict their reading fluency. This research is important for advancing personalized education by using AI to provide real-time assessments of students’ reading fluency. It addresses a critical need in elementary education to support students in developing reading skills, which are fundamental for academic success and comprehension. By integrating these assessments into educational technologies, the project aims to offer timely feedback and personalized interventions to improve reading outcomes for young learners.
Responsibilities
In this project, I developed features for an AI-based stealth assessment of oral reading fluency (ORF) by analyzing audio recordings of students’ reading practice. I created features spanning phonetic, lexical, syntactic, and semantic levels to evaluate reading fluency, and applied machine learning techniques, such as random forests and neural networks, to assess their importance in predicting reading fluency. This involved using advanced algorithms to derive insights from complex data, enhancing my proficiency in feature engineering and machine learning. Additionally, I learned about stealth assessment methods and their application in personalized educational technologies. I am preparing to communicate these findings through a research paper and presentation at a major educational conference.