Backtranslation Feedback Improves User Confidence in MT, Not Quality

TL;DR

This study explores how feedback mechanisms like backtranslation, quality estimation, and paraphrasing influence user confidence and perceived quality in outbound machine translation from English to Czech and Estonian.

Contribution

It introduces an experiment analyzing the effects of various feedback modules on user confidence and translation quality in outbound translation tasks.

Findings

Backward translation increases user confidence but not actual translation quality.

Feedback modules influence user perception differently depending on translation system quality.

User confidence can be improved independently of objective translation accuracy.

Abstract

Translating text into a language unknown to the text's author, dubbed outbound translation, is a modern need for which the user experience has significant room for improvement, beyond the basic machine translation facility. We demonstrate this by showing three ways in which user confidence in the outbound translation, as well as its overall final quality, can be affected: backward translation, quality estimation (with alignment) and source paraphrasing. In this paper, we describe an experiment on outbound translation from English to Czech and Estonian. We examine the effects of each proposed feedback module and further focus on how the quality of machine translation systems influence these findings and the user perception of success. We show that backward translation feedback has a mixed effect on the whole process: it increases user confidence in the produced translation, but not the objective quality.