Approaching English-Polish Machine Translation Quality Assessment with Neural-based Methods

TL;DR

This paper explores neural-based methods for evaluating English-Polish machine translation quality, utilizing pre-trained language models to improve assessment accuracy, achieving top-tier rankings in a competitive evaluation.

Contribution

It introduces neural-based approaches using pre-trained models for translation quality assessment, advancing the state-of-the-art in both blind and nonblind evaluation settings.

Findings

Ranked second in nonblind evaluation

Ranked third in blind evaluation

Demonstrated effectiveness of neural models in MT quality assessment

Abstract

This paper presents our contribution to the PolEval 2021 Task 2: Evaluation of translation quality assessment metrics. We describe experiments with pre-trained language models and state-of-the-art frameworks for translation quality assessment in both nonblind and blind versions of the task. Our solutions ranked second in the nonblind version and third in the blind version.