What is Language Weavers method of automated translation, and how is it different?
Traditionally, automated (machine) translation techniques have been rule-based. This required many linguists, spending many months, to develop systems that correlated the rules of syntax and grammar from one language to another. The problem with this method is that there are too many exceptions to the rules and the translation output from these systems sounded unnatural and was often incorrect. Language Weaver pioneered a new approach to automated translation called statistical machine translation software, which fundamentally changes the nature of what can be done with automated translation. Language Weaver’s approach is statistical – that is, the translation engines are built using pattern recognition and probability statistics. This technique trains the translation engine with pre-translated parallel documents, essentially allowing the computer to learn from existing human translations. For those familiar with other translation technologies, Language Weaver’s translation software is
