Can I read 2022.amta Research.24 on EtoBox?

2022.amta Research.24 by Mohamed Gamal is a document available to read on EtoBox.

What is 2022.amta Research.24 about?

This document presents a new metric, XSTS, aimed at improving human evaluation consistency in machine translation across various language pairs by focusing on semantic equivalence. It introduces a calibration method to address variability in evaluator standards, demonstrating higher inter-annotator agreement and better correlation with automatic scores through large-scale studies. The findings suggest that XSTS and the calibration approach enhance the reliability of quality assessments in machine translatio

Author

Mohamed Gamal

Language

EN