CUNI System for the WMT19 Robustness Task
This work addresses robustness in machine translation for noisy inputs, but it is incremental as it builds on existing systems and tasks.
The paper tackled the WMT19 Robustness Task by evaluating a Transformer-based system, showing it was far more robust to noisy input than an LSTM baseline, and improved performance through fine-tuning on noisy data without harming news translation quality.
We present our submission to the WMT19 Robustness Task. Our baseline system is the Charles University (CUNI) Transformer system trained for the WMT18 shared task on News Translation. Quantitative results show that the CUNI Transformer system is already far more robust to noisy input than the LSTM-based baseline provided by the task organizers. We further improved the performance of our model by fine-tuning on the in-domain noisy data without influencing the translation quality on the news domain.