ABSTRACT
Natural language processing (NLP) is an academic field for developing and evaluating technologies for recognising and generating human languages. For instance, computational methods that have so far helped people working in the translation production process, such as translation memory and bilingual concordancer, are well-known outcomes of NLP. Unlike the theoretical methodologies studied for decades, recent advances, statistical and neural approaches, exploit an abundant amount of electronically accumulated textual data, such as those on the Web. In the last couple of years, owing to significant computational resources, neural machine translation (NMT) has replaced classical statistical approaches and recent research suggests that NMT can achieve parity with human translations produced by professionals. This chapter explains the mechanism of NMT, points out its limitations, and discusses its potential extensions in line with effective human-computer collaboration. Besides NMT, this chapter also introduces NLP tools that are built in computer-assisted translation frameworks, followed by a discussion on the requirements of usefulness.
