Header Image

Alan Melby

Speaker

Alan Melby has taught translation for many years at Brigham Young University. After working on a machine translation project for ten years, he studied human translation and became a certified French-to-English translator.

Alan Melby has taught translation for many years at Brigham Young University. After working on a machine translation project for ten years, he studied human translation and became a certified French-to-English translator. In 2014 he was elected by the International Federation of Translators (FIT) Council at the Congress in Berlin. 

Bells for Machine Translation (MT) are not ringing yet 

We are now in the third generation of machine translation systems. The first was rule-based machine translation (RBMT); the second was statistical machine translation (SMT); and the current generation is based on “neural” networks (Neural machine translation, NMT). Claims are being made that the need for professional human translators will soon disappear, because machine translation has achieved “parity” with human translation. This presentation (1) introduces Multi-dimensional Metrics (MQM) as a tool in evaluating claims of parity; (2) describes a practical method of comparing human translation providers and machine translation systems; and (3) suggests that collaboration between human translators and machine translation system developers is more productive than competition.