ÃÈÃÃÉçÇø

Currently, despite most students having the potential to read at or above grade level, only 30% of U.S. 4th grade students meet reading proficiency standards. Students with less economic resources, those with disabilities, and students of color have been especially vulnerable to adverse conditions that thwart efforts to reach reading potential in early years (with lasting COVID-19 effects contributing some of those conditions). Despite empirical research that demonstrates we know what and how to teach reading, these findings do not yet translate to classroom settings. This year’s SCOPE team, like the 2024-2025 team will join a large interdisciplinary research team and contribute key components to proof-of-concept instructional games for K-3rd grade students developing reading skills. The SCOPE team will explore getting games as close as possible to "teacher level" accuracy levels for the task of determining whether words read aloud by students of different backgrounds (such as English as a second language learners) were correct. Additionally, SCOPE students will optimize large language models for generating in-game activities for individual readers that are appropriate for their reading level (both dynamically in one reading session and across multiple sessions). Some optimizations will explore the feasibility, desirability, and viability of recognizing and responding to variations in speech across Mainstream American English (MAE), different regional dialects, diglossia (e.g., Spanish/English hybrids or creoles), or culture-specific vernaculars. 

Faculty Advisor:

• Amon Millner

Team Members: