ABSTRACT
Chapter 3 takes neural machine translation (NMT), the latest form of MT, as a starting point. It is discussed as a TEnT in CAT tools. You will learn about the different ways of integrating NMT in the CAT tool, the required APIs and plugins, the interaction between NMT and the translation memory (TM), and how NMT becomes adaptive. Integrated and adaptive MT features enable post-editing and quality control on the fly. PEMT should not be necessary after the translation is completed. Quality is a fluid concept. In the CAT tool, it is difficult to measure MT quality because it is a process and not a product. The industry has designed frameworks and metrics to measure the impact on translation quality of integrated or adaptive MT. The project assignment focuses on trial and evaluation of MT as an integrated tool. Desktop CAT programs can be used for the assignment depending on the availability of MT search engines, APIs and plugins. Translation management systems (full discussion in Chapter 7) are a suitable alternative as they fully integrate TM and MT. The assignment is not about measuring machine translated texts, but about MT as a supportive TEnT in the CAT tool.
