Google puts neural machine translation into production

Google switched Google Translate’s Chinese-to-English pair to GNMT, an end-to-end deep LSTM system described in a companion paper by Yonghui Wu and 30 co-authors. Google said it cut translation errors by 55 to 85 percent on several language pairs.

Why it mattered A switch handling about 18 million translations a day showed that deep learning had become fast and reliable enough for production work rather than demonstrations.