From 607285eac9041f4af6460422c63bf7304eb17ea8 Mon Sep 17 00:00:00 2001 From: Akshaya4225 Date: Sat, 28 Mar 2026 11:36:56 +0530 Subject: [PATCH] Enhance project_AutoEIT.md with dashboard details Added a section on Metrics Dashboard and Results Visualization for the AutoEIT project. --- _gsocprojects/2026/project_AutoEIT.md | 24 +++++++++++++++++++++++- 1 file changed, 23 insertions(+), 1 deletion(-) diff --git a/_gsocprojects/2026/project_AutoEIT.md b/_gsocprojects/2026/project_AutoEIT.md index 2b898713..d922bdf8 100644 --- a/_gsocprojects/2026/project_AutoEIT.md +++ b/_gsocprojects/2026/project_AutoEIT.md @@ -3,7 +3,29 @@ project: AutoEIT layout: default logo: AutoEIT.png description: | - 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. + 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. + +--- + +## Metrics Dashboard and Results Visualization + +AutoEIT can include a dashboard to visualize transcription and scoring results: + +1. **Score Distribution** + - Visualize the range of scores for all participants + - Identify patterns in common errors + +2. **Transcription Accuracy Metrics** + - Word Error Rate (WER) per audio file + - Highlight low-accuracy transcriptions for review + +3. **Real-Time Updates** + - Display ongoing processing results as they are scored + - Color-coded alerts for scores below threshold + +4. **Export Options** + - Export results as CSV, JSON, or PDF + - Shareable reports for researchers or instructors ---