DFKI-LT - Multi-source Neural Automatic Post-Editing: FBK's participation in the WMT 2017 APE shared task

Rajen Chatterjee, M. Amin Farajian, Matteo Negri, Marco Turchi, Ankit Srivastava, Santanu Pal
Multi-source Neural Automatic Post-Editing: FBK's participation in the WMT 2017 APE shared task
2 Second Conference on Machine Translation, Pages 630-638, Copenhagen, Denmark, Association for Computational Linguistics, 9/2017
 
Previous phrase-based approaches to Au- tomatic Post-editing (APE) have shown that the dependency of MT errors from the source sentence can be exploited by jointly learning from source and target informa- tion. By integrating this notion in a neu- ral approach to the problem, we present the multi-source neural machine transla- tion (NMT) system submitted by FBK to the WMT 2017 APE shared task. Our system implements multi-source NMT in a weighted ensemble of 8 models. The n-best hypotheses produced by this en- semble are further re-ranked using fea- tures based on the edit distance between the original MT output and each APE hy- pothesis, as well as other statistical models (n-gram language model and operation se- quence model). This solution resulted in the best system submission for this round of the APE shared task for both en-de and de-en language directions. For the for- mer language direction, our primary sub- mission improves over the MT baseline up to -4.9 TER and +7.6 BLEU points. For the latter, where the higher quality of the original MT output reduces the room for improvement, the gains are lower but still significant (-0.25 TER and +0.3 BLEU).
 
Files: BibTeX, WMT73.pdf, WMT73.pdf