Automated Scoring System for Elicited Imitation Task Responses

AutoEIT is an NLP system for automating scoring of the Elicited Imitation Task (EIT), a language proficiency assessment where participants repeat spoken sentences. The project builds an end-to-end pipeline that transcribes audio and assigns rubric-based scores (0 to 4) by comparing responses with target sentences. It combines speech processing, linguistic rules, semantic similarity, and LLM-based evaluation. The goal is to produce reliable, scalable scoring aligned with human raters.