This proposal focuses on creating a reliable, standardized, fully automated scoring system for the Spanish Elicited Imitation Task (EIT), using the transcriptions generated from learner audio. The EIT scoring process is currently labor-intensive and fully dependent on trained human raters. Although large language models can assist with scoring, they often produce inconsistent results, awarding different scores to the exact same sentence across sessions or prompts. This lack of standardization makes automated scoring using existing tools unsuitable for research purposes.
The goal of this project is to eliminate scoring variability by designing a system that applies the EIT scoring rubric in a consistent, rule-driven, and reproducible way. This tool will support research involving second/additional language learners by providing high-quality, standardized, automated scoring for large datasets. By developing both the core scoring logic and a web-based interface, the project delivers a practical tool that will significantly streamline research on second and additional language learning.
Total project length: 175 hours
Python, Pytorch or Tensorflow, and some previous experience in Machine Learning.
Medium
Please DO NOT contact mentors directly by email. Instead, please email human-ai@cern.ch with Project Title and include your CV and test results. The mentors will then get in touch with you.