The UN Parallel Corpus Annotated for Translation Direction

TL;DR

This paper presents a method to distinguish translated from original texts in the UN corpus with high accuracy, by modeling it as a classification problem using various feature extraction techniques, and provides a publicly available annotated corpus.

Contribution

It introduces a new annotated parallel corpus for translation direction and compares multiple classification methods for distinguishing translated and original texts.

Findings

Achieved up to 95% classification accuracy

Compared different feature extraction methods

Provided publicly available annotated corpus

Abstract

This work distinguishes between translated and original text in the UN protocol corpus. By modeling the problem as classification problem, we can achieve up to 95% classification accuracy. We begin by deriving a parallel corpus for different language-pairs annotated for translation direction, and then classify the data by using various feature extraction methods. We compare the different methods as well as the ability to distinguish between translated and original texts in the different languages. The annotated corpus is publicly available.