Dual Conditional Cross-Entropy Filtering of Noisy Parallel Corpora
This work addresses the challenge of improving machine translation quality by effectively filtering noisy parallel data, which is incremental as it builds on existing filtering techniques with a novel scoring approach.
The paper tackles the problem of filtering noisy parallel corpora by introducing dual conditional cross-entropy filtering, which computes cross-entropy scores using two inverse translation models and penalizes divergent scores. The method achieved higher BLEU scores than models trained on clean WMT data and ranked highest in the WMT2018 shared task, scoring top in three out of four subtasks.
In this work we introduce dual conditional cross-entropy filtering for noisy parallel data. For each sentence pair of the noisy parallel corpus we compute cross-entropy scores according to two inverse translation models trained on clean data. We penalize divergent cross-entropies and weigh the penalty by the cross-entropy average of both models. Sorting or thresholding according to these scores results in better subsets of parallel data. We achieve higher BLEU scores with models trained on parallel data filtered only from Paracrawl than with models trained on clean WMT data. We further evaluate our method in the context of the WMT2018 shared task on parallel corpus filtering and achieve the overall highest ranking scores of the shared task, scoring top in three out of four subtasks.