Train, Sort, Explain: Learning to Diagnose Translation Models

Robert Schwarzenberg, David Harbecke, Vivien Macketanz, Eleftherios Avramidis, Sebastian Möller


Abstract
Evaluating translation models is a trade-off between effort and detail. On the one end of the spectrum there are automatic count-based methods such as BLEU, on the other end linguistic evaluations by humans, which arguably are more informative but also require a disproportionately high effort. To narrow the spectrum, we propose a general approach on how to automatically expose systematic differences between human and machine translations to human experts. Inspired by adversarial settings, we train a neural text classifier to distinguish human from machine translations. A classifier that performs and generalizes well after training should recognize systematic differences between the two classes, which we uncover with neural explainability methods. Our proof-of-concept implementation, DiaMaT, is open source. Applied to a dataset translated by a state-of-the-art neural Transformer model, DiaMaT achieves a classification accuracy of 75% and exposes meaningful differences between humans and the Transformer, amidst the current discussion about human parity.
Anthology ID:
N19-4006
Volume:
Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics (Demonstrations)
Month:
June
Year:
2019
Address:
Minneapolis, Minnesota
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
29–34
Language:
URL:
https://aclanthology.org/N19-4006
DOI:
10.18653/v1/N19-4006
Bibkey:
Copy Citation:
PDF:
https://aclanthology.org/N19-4006.pdf