Work

Long-context Reference-based MT Quality Estimation

Year: 2025

Type: conference-paper

Abstract: In this paper, we present our submission to the Tenth Conference on Machine Translation (WMT25) Shared Task on Automated Translation Quality Evaluation.Our systems are built upon the COMET framework and trained to predict segment-level Error Span Annotation (ESA) scores using augmented long-context data.To construct long-context training data, we concatenate in-domain, human-annotated sentences and compute a weighted average of their scores.We integrate multiple human judgment datasets (MQM, SQM, and DA) by normalising their scales and train multilingual regression models to predict quality scores from the source, hypothesis, and reference translations.Experimental results show that incorporating long-context information improves correlations with human judgments compared to models trained only on short segments.

Source: Conference on Machine Translation (WMT)

DOI: 10.18653/v1/2025.wmt-1.64

Authors: Sami Haq, Chinonso Osuji, Sheila Castilho, Brian Davis, Thiago Castro Ferreira

Institutions: Dublin City University, ADAPT - Centre for Digital Content Technology, Universidade Federal Fluminense

Language: English

Topic: Natural Language Processing Techniques

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Funders: Science Foundation Ireland, Dublin City University, Research Ireland