Factuality Detection using Machine Translation -- a Use Case for German Clinical Text

TL;DR

This paper proposes a method for factuality detection in German clinical text by translating English data into German to train a transformer-based model, addressing data sharing constraints.

Contribution

It introduces a novel approach using machine translation to leverage English data for training factuality detection models in German clinical text.

Findings

Effective translation-based training improves factuality detection accuracy.

Addresses data sharing limitations in clinical NLP.

Demonstrates feasibility of cross-lingual model training for clinical tasks.

Abstract

Factuality can play an important role when automatically processing clinical text, as it makes a difference if particular symptoms are explicitly not present, possibly present, not mentioned, or affirmed. In most cases, a sufficient number of examples is necessary to handle such phenomena in a supervised machine learning setting. However, as clinical text might contain sensitive information, data cannot be easily shared. In the context of factuality detection, this work presents a simple solution using machine translation to translate English data to German to train a transformer-based factuality detection model.