Lemmatic Machine Translation
Stephen Soderland, Christopher Lim, Mausam Mausam, Bo Qin, Oren Etzioni, Jonathan Robert Pool · 2009
Statistical MT is limited by reliance on large parallel corpora. We propose Lemmatic MT, a new paradigm that extends MT to a far broader set of languages, but requires substantial manual encoding effort. We present PANLINGUAL TRANSLATOR, a prototype Lemmatic MT system with high translation adequacy on 59 % to 99 % of sentences (average 84%) on a sample of 6 language pairs that Google Translate (GT) handles. GT ranged from 34 % to 93%, average 65%. PANLINGUAL TRANSLATOR also had high translation adequacy on 27 % to 82 % of sentences (average 62%) from a sample of 5 language pairs not handled by GT.