AutoEIT is an applied machine learning project focused on automating the scoring of the Elicited Imitation Task (EIT), a research tool used to measure global language proficiency. The EIT is widely respected and is available for free in several languages, but the current workflow requires manual audio transcription and human scoring—slow, tedious, and error‑prone. This project aims to build an end‑to‑end system that will: Process raw audio files, perform accurate voice‑to‑text transcription, and automatically evaluate responses using a standardized scoring rubric.
Audio-to-text transcription for second/additional language learner data
Automated scoring system for elicited imitation task responses