Neural Machine Translation
A machine translation approach based on neural networks that uses deep learning to perform end-to-end translation from a source language to a target language. It represents the dominant paradigm in contemporary machine translation.
Timeline (last 90 days)
The attention mechanism was first introduced into the field of neural machine translation around 2014, to address the problem of information loss in long sequences.
On-device offline translation models are typically based on lightweight neural machine translation architectures (such as quantized Transformer variants), compressed to tens to hundreds of MB in size, and capable of real-time inference on a CPU or local NPU.
The introduction of neural network architectures in 2014 enabled end-to-end translation model training, and Bahdanau et al. proposed attention-based sequence-to-sequence models, ushering in the era of Neural Machine Translation (NMT).
There is a MAP decoding paradox in neural machine translation: the sequence with the highest probability is not equivalent to the sequence with the best translation quality, and degenerate phenomena such as empty or excessively short sequences may even occur.
All Facts (6)
Google Cloud Translation API是谷歌基于神经机器翻译（NMT）技术构建的云端翻译服务，支持100余种语言
90%UnverifiedThe quality of AI-translated technical documentation has reached a level where it can be read fluently, greatly lowering the learning barrier for non-English-speaking developers.
60%UnverifiedThe attention mechanism was first introduced into the field of neural machine translation around 2014, to address the problem of information loss in long sequences.
50%UnverifiedOn-device offline translation models are typically based on lightweight neural machine translation architectures (such as quantized Transformer variants), compressed to tens to hundreds of MB in size, and capable of real-time inference on a CPU or local NPU.
50%UnverifiedThe introduction of neural network architectures in 2014 enabled end-to-end translation model training, and Bahdanau et al. proposed attention-based sequence-to-sequence models, ushering in the era of Neural Machine Translation (NMT).
50%UnverifiedThere is a MAP decoding paradox in neural machine translation: the sequence with the highest probability is not equivalent to the sequence with the best translation quality, and degenerate phenomena such as empty or excessively short sequences may even occur.
50%