KIRA screening system improves early reading checks with phoneme feedback
KIRA: A Full-Stack Reading Screening System with Dual-Pathway Analysis for Children
Computation and LanguageMachine Learning
Summary
Many young children struggle with reading, and identifying those who need extra help early on is important. The authors created KIRA, a tool that listens to kids reading aloud and checks both whole words and the smaller sounds inside words. By adapting speech recognition models specifically for children's voices, KIRA reduces errors in detecting words and sounds. It also provides teachers with easy-to-understand scores and detailed feedback to help guide reading support.
What this means in practice
- •For school it administrators: Deploy a classroom-ready tool that screens K-2 students’ reading with detailed fluency scores and phoneme-level feedback to support targeted interventions.
- •For speech therapy clinics: Use detailed phoneme error detection from child-adapted speech models to aid diagnosis and monitor progress in young children’s speech and reading skills.
Authors
Maggie Dong, Rohit Mekkoth, Zixuan Wang, Cheol Jun Cho, Jiachen Lian, Gopala Anumanchipalli
Abstract
Early-grade reading screening is critical for identifying children at risk of reading difficulty. Existing AI-based screeners are often limited in the scope of their analyses and evaluation metrics, or are proprietary systems that lack transparent model specifications and evaluation. We present KIRA, a dual-pathway screening system that pairs word-level fluency scoring with detailed phoneme-level error detection. We adapt pretrained word and phoneme recognizers to children's speech, reducing word error rate from 13.88% to 8.84% and phoneme error rate from 41.66% to 23.00%. Powered by these child-adapted models, we implement Word Reading (WRE) and Oral Reading Fluency (ORF), two foundational tasks in K-2 screening. The system integrates these recognizers with a web-based teacher-facing interface presenting fluency scores alongside phoneme feedback for potential targeted intervention. KIRA brings modern child-speech recognition into a full-stack, deployable, classroom-ready reading screener.