Vol. 6:2 (2017) ► pp.291–309
Making sense of neural machine translation
The last few years have witnessed a surge in the interest of a new machine translation paradigm: neural machine translation (NMT). Neural machine translation is starting to displace its corpus-based predecessor, statistical machine translation (SMT). In this paper, I introduce NMT, and explain in detail, without the mathematical complexity, how neural machine translation systems work, how they are trained, and their main differences with SMT systems. The paper will try to decipher NMT jargon such as “distributed representations”, “deep learning”, “word embeddings”, “vectors”, “layers”, “weights”, “encoder”, “decoder”, and “attention”, and build upon these concepts, so that individual translators and professionals working for the translation industry as well as students and academics in translation studies can make sense of this new technology and know what to expect from it. Aspects such as how NMT output differs from SMT, and the hardware and software requirements of NMT, both at training time and at run time, on the translation industry, will be discussed.
- 2.What is neural machine translation and how does it work?
- 2.1Neural machine translation is corpus-based machine translation
- 2.2Neural machine translation uses neural networks
- 2.2.1Neural units or neurons
- 2.2.2Grouping units into layers to learn distributed representations
- 2.3How does neural machine translation work?
- 2.3.2Machine translation as predicting the next word
- 2.3.3Representations for words and for longer segments of text
- 2.4Extensions and alternative neural machine translation architectures
- 2.4.2“Convolutional” neural machine translation
- 2.4.3Doing away with recursion and convolution: is attention all you need?
- 2.5Main differences between neural and statistical machine translation
- 3.What can translators expect from neural machine translation?
- 3.1High computational requirements
- 3.2A different kind of output
- 3.3Is neural machine translation better than statistical machine translation?
- 3.3.1Automatic evaluation
- 3.3.2Subjective evaluation
- 3.3.3Measuring post-editing effort and productivity
- 4.Concluding remarks
This article is currently available as a sample article.
Cited by 44 other publications
This list is based on CrossRef data as of 17 may 2023. Please note that it may not be complete. Sources presented here have been supplied by the respective publishers. Any errors therein should be reported to them.