CLJun 12, 2017
Six Challenges for Neural Machine Translation
arXiv:1706.03872v11549 citations
Originality Synthesis-oriented
AI Analysis
This work addresses critical bottlenecks for researchers and practitioners in machine translation, though it is incremental as it builds on existing methods.
The paper identifies six key challenges in neural machine translation, including domain mismatch and rare words, and demonstrates both deficiencies and improvements compared to phrase-based statistical machine translation.
We explore six challenges for neural machine translation: domain mismatch, amount of training data, rare words, long sentences, word alignment, and beam search. We show both deficiencies and improvements over the quality of phrase-based statistical machine translation.