Physician Detection of Clinical Harm in Machine Translation: Quality Estimation Aids in Reliance and Backtranslation Identifies Critical Errors

Nikita Mehandru, Sweta Agrawal, Yimin Xiao, Ge Gao, Elaine C Khoong, Marine Carpuat, Niloufar Salehi

Submission Type: Regular Long Paper

Submission Track: Human-Centered NLP

Keywords: medical machine translation, clinical harm, human-centered NLP

TL;DR: We conduct a human study with physicians to test how certain interventions affect their ability to detect machine translation errors in high-stakes medical settings.

Abstract: A major challenge in the practical use of Machine Translation (MT) is that users lack information on translation quality to make informed decisions about how to rely on outputs. Progress in quality estimation research provides techniques to automatically assess MT quality, but these techniques have primarily been evaluated in vitro by comparison against human judgments outside of a specific context of use. This paper evaluates quality estimation feedback in vivo with a human study in realistic high-stakes medical settings. Using Emergency Department discharge instructions, we study how interventions based on quality estimation versus backtranslation assist physicians in deciding whether to show MT outputs to a patient. We find that quality estimation improves appropriate reliance on MT, but backtranslation helps physicians detect more clinically harmful errors that QE alone often misses.

Submission Number: 4240